Hubbry Logo
Software patentSoftware patentMain
Open search
Software patent
Community hub
Software patent
logo
7 pages, 0 posts
0 subscribers
Be the first to start a discussion here.
Be the first to start a discussion here.
Software patent
Software patent
from Wikipedia

A software patent is a patent on a piece of software, such as a computer program, library, user interface, or algorithm. The validity of these patents can be difficult to evaluate, as software is often at once a product of engineering, something typically eligible for patents, and an abstract concept, which is typically not. This gray area, along with the difficulty of patent evaluation for intangible, technical works such as libraries and algorithms, makes software patents a frequent subject of controversy and litigation.

Different jurisdictions have radically different policies concerning software patents, including a blanket ban, no restrictions, or attempts to distinguish between purely mathematical constructs and "embodiments" of these constructs. For example, an algorithm itself may be judged unpatentable, but its use in software judged patentable.

Background

[edit]

A patent is a set of exclusionary rights granted by a state to a patent holder for a limited period of time, usually 20 years. These rights are granted to patent applicants in exchange for their disclosure of the inventions. Once a patent is granted in a given country, no person may make, use, sell or import/export the claimed invention in that country without the permission of the patent holder. Permission, where granted, is typically in the form of a license which conditions are set by the patent owner: it may be free or in return for a royalty payment or lump sum fee.

Patents are territorial in nature. To obtain a patent, inventors must file patent applications in each and every country in which they want a patent. For example, separate applications must be filed in Japan, China, the United States and India if the applicant wishes to obtain patents in those countries. However, some regional offices exist, such as the European Patent Office (EPO), which act as supranational bodies with the power to grant patents which can then be brought into effect in the member states, and an international procedure also exists for filing a single international application under the Patent Cooperation Treaty (PCT), which can then give rise to patent protection in most countries.

These different countries and regional offices have different standards for granting patents. This is particularly true of software or computer-implemented inventions, especially where the software is implementing a business method.

Early example of a software patent

[edit]

On 21 May 1962, a British patent application entitled "A Computer Arranged for the Automatic Solution of Linear Programming Problems" was filed.[1] The invention was concerned with efficient memory management for the simplex algorithm, and could be implemented by purely software means. The patent struggled to establish that it represented a 'vendible product'. "The focus of attention shifted to look at the relationship between the [unpatentable] computer program and the [potentially patentable] programmed computer".[2] The patent was granted on August 17, 1966, and seems to be one of the first software patents, establishing the principle that the computer program itself was unpatentable and therefore covered by copyright law, while the computer program embedded in hardware was potentially patentable.[3][4]

Jurisdictions

[edit]

Most countries place some limits on the patenting of inventions involving software, but there is no one legal definition of a software patent. For example, U.S. patent law excludes "abstract ideas", and this has been used to refuse some patents involving software. In Europe, "computer programs as such" are excluded from patentability, thus European Patent Office policy is consequently that a program for a computer is not patentable if it does not have the potential to cause a "technical effect" which is by now understood as a material effect (a "transformation of nature").[5] Substantive law regarding the patentability of software and computer-implemented inventions, and case law interpreting the legal provisions, are different under different jurisdictions.

Software patents under multilateral treaties:

Software patents under national laws:

Australia

[edit]

In Australia, there is no particular exclusion for patents relating to software. The subject matter of an invention is patentable in Australia, if it is a manner of manufacture within the meaning of section 6 of the Statute of Monopolies.[6] The High Court of Australia has refrained from ruling on the precise definition of manner of manufacture stating that any such attempt is bound to fail for the policy reason of encouraging national development in fields that may be unpredictable.[7] In assessing whether an invention is a manner of manufacture, the High Court has relied on the inquiry of whether the subject of the claims defining the invention has as its end result an artificially created state of affairs.[7]

In a decision of the Federal Court of Australia, on the patentability of an improved method of representing curved images in computer graphics displays, it was held that the application of selected mathematical methods to computers may involve steps which are foreign to the normal use of computers and hence amount to a manner of manufacture.[8] In another unanimous decision by the Full Federal Court of Australia, an invention for methods of storing and retrieving Chinese characters to perform word processing was held to be an artificially created state of affairs and consequently within the concept of a manner of manufacture.[9]

Nevertheless, in a recent decision on the patentability of a computer implemented method of generating an index based on selection and weighing of data based on certain criterion, the Full Federal Court of Australia reaffirmed that mere methods, schemes and plans are not manners of manufacture.[10] The Full Court went on to hold that the use of a computer to implement a scheme did not contribute to the invention or the artificial effect of the invention.[10] The subject matter of the invention was held to be an abstract idea and not a manner of manufacture within the meaning of the term in the Patents Act. The same Full Federal Court in another decision regarding the patentability of an invention regarding a method and system for assessing an individual's competency in relation to certain criterion, reiterated that a business method or mere scheme were per se are not patentable.[11]

In principle, computer software is still a valid patentable subject matter in Australia. But, in circumstances where patents have been sought over software to merely implement abstract ideas or business methods, the courts and the Commissioner of Patents have resisted granting patent protection to such applications both as a matter of statutory interpretation and policy.

Canada

[edit]

In Canada, courts have held that the use of a computer alone neither lends, nor reduces patentability of an invention. However, it is the position of the Canadian Patent Office that where a computer is an "essential element" of a patent's claims, the claimed invention is generally patentable subject matter.[12]

China

[edit]

In China, the starting time of software patent is relatively late. Before 2006, software patents were basically not granted, and software and hardware had to be combined when applying for a patent. With the development of network technology and software technology, China's patent examination system has been constantly updated. Recently, the design idea of the software itself has been allowed to apply for patent separately, instead of requiring to be combined with hardware. However, software patent writing requirements are relatively high.

Software patents can be written as either a product or a method, depending on the standards of review. However, no matter what form it is written in, it is difficult to highlight the creativity of the scheme, which requires specific case analysis.

Software that can be patented mainly includes (but is not limited to):

(1) Industrial control software, such as controlling the movement of mechanical equipment;

(2) Software to improve the internal performance of the computer, such as a software can improve the virtual memory of the computer;

(3) External technical data processing software, such as digital camera image processing software.

It is fair to say that a considerable proportion of software belongs to category (3).

The patent protection measures can be seen in the patent law and the regulations on the protection of computer software.

Europe

[edit]

Within European Union member states, the EPO and other national patent offices have issued many patents for inventions involving software since the European Patent Convention (EPC) came into force in the late 1970s. Article 52 EPC excludes "programs for computers" from patentability (Art. 52(2)) to the extent that a patent application relates to a computer program "as such" (Art. 52(3)). This has been interpreted to mean that any invention that makes a non-obvious "technical contribution" or solves a "technical problem" in a non-obvious way is patentable even if that technical problem is solved by running a computer program.[13] When the EPO examines a patent application with questionable subject matter eligibility, their approach is to simply[dubiousdiscuss] disregard any ineligible portions or aspects and evaluate the rest.[14] This is notably different from the U.S. approach (see below).

Computer-implemented inventions that only solve a business problem using a computer, rather than a technical problem, are considered unpatentable as lacking an inventive step (see T 258/03). Nevertheless, the fact that an invention is useful in business does not mean it is not patentable if it also solves a technical problem.

A summary of the developments concerning patentability of computer programs under the European Patent Convention is given in (see G 3/08) as a response of the Enlarged Board of Appeal to questions filed by the President of the European Patent Office according to Article 112(1)(b) EPC.[needs update]

Concerns have been raised by free software campaigners, such as the Free Software Foundation, that the Unified Patent Court will be much more open to patents generally and software patents in particular.[15]

Germany

[edit]

In April 2013, the German Parliament adopted a joint motion "against the growing trend of patent offices to grant patents on software programs".[16]

United Kingdom

[edit]

United Kingdom patent law is interpreted to have the same effect as the European Patent Convention such that "programs for computers" are excluded from patentability to the extent that a patent application relates to a computer program "as such". Current case law in the UK states that an (alleged) invention will only be regarded as an invention if it provides a contribution that is not excluded and that is also technical. A computer program implementing a business process is therefore not an invention, but a computer program implementing an industrial process may well be.

India

[edit]

In India, a clause to include software patents was quashed by the Indian Parliament in April 2005.[17] However, following publication of the new guidelines on the examination of computer-related inventions on 19 February 2016, the Office of the Controller General of Patents, Designs and Trade marks accepts applications for software patents, as long as the software is claimed in conjunction with a novel hardware.[18] On 30 June 2017, revised guidelines on the examination of computer related inventions were published. This 2017 guidelines provides clarity on patentability of software invention in India, i.e., the claimed computer-related invention needs to be ascertained whether it is of a technical nature involving technical advancement as compared to the existing knowledge or having economic significance or both, and is not subject to exclusion under Section 3 of the Patents Act.[19] In 2019, the Court observed,

In today’s digital world, when most inventions are based on computer programs, it would be retrograde to argue that all such inventions would not be patentable. Innovation in the field of artificial intelligence, blockchain technologies and other digital products would be based on computer programs, however the same would not become nonpatentable inventions – simply for that reason. It is rare to see a product which is not based on a computer program. Whether they are cars and other automobiles, microwave ovens, washing machines, refrigerators, they all have some sort of computer programs in-built in them. Thus, the effect that such programs produce including in digital and electronic products is crucial in determining the test of patentability.

Patent applications in these fields would have to be examined to see if they result in a “technical contribution”, it added. Further elaborating on the usage of the term ‘per se’ in Section 3(k), the Court said,

The words ‘per se’ were incorporated so as to ensure that genuine inventions which are developed, based on computer programs are not refused patents.[20]

With respect to the term per se, the joint parliamentary committee had expressed the following view:

In the new proposed clause (k) the words: “per se” have been inserted. This change has been proposed because sometimes the computer programme may include certain other things, ancillary thereto or developed thereon. The intention here is not to reject them for grant of patent if they are inventions. However, the computer programs as such are not intended to be granted patent. This amendment has been proposed to clarify the purpose.[21]

Japan

[edit]

Software-related inventions are patentable. To qualify as an invention, however, there must be "a creation of technical ideas utilizing a law of nature"[22] although this requirement is typically met by "concretely realising the information processing performed by the software by using hardware resources".[23] Software-related inventions may be considered obvious if they involve the application of an operation known in other fields, the addition of a commonly known means or replacement by equivalent, the implementation in software of functions which were previously performed by hardware, or the systematisation of known human transactions.[24]

In 1999, the allowance rate for business method patents at the Japan Patent Office (JPO) reached an all-time high of roughly 35 percent. Subsequently, the JPO experienced a surge in business method patent filings. This surge was met with a dramatic decrease in the average grant rate of business method patents during the following six years; it lingered around 8 percent between 2003 and 2006 (8 percent is extremely low in comparison to the average of 50 percent across all technical fields). A report from 2012 found that the average grant rate since 2006 for business method patents has risen to the current rate of roughly 25 percent.[25]

New Zealand

[edit]

In New Zealand computer programs are excluded from patentability under the Patents Act 2013,[26] but guidelines permitting embedded software were added since the initial Patents Bill.[27] From 2013 computer programs 'as such' are excluded from patentability. The as such wording rules out only those software based patents where novelty lies solely in the software. Similar to Europe.[28]

Philippines

[edit]

In the Philippines, "schemes, rules and methods of performing mental acts, playing games or doing business, and programs for computers" are non-patentable inventions under Sec. 22.2 of Republic Act No. 8293, otherwise known as the "Intellectual Property Code of the Philippines".

Russian Federation

[edit]

In the Russian Federation according to article #1350 of the Civil Code of the Russian Federation the following do not qualify as inventions:

  1. discoveries;
  2. scientific theories and mathematical methods;
  3. solutions concerning only the appearance of products and aimed at meeting the aesthetic needs;
  4. the rules and methods of games, intellectual or economic activities;
  5. computer programs;
  6. solutions consisting only in the presentation of information.

However, the article provides for that the patentability of these objects is excluded only in the case when the application for the grant of a patent for an invention concerns these objects as such.

South Africa

[edit]

In South Africa, "a program for a computer" is excluded from recognition as an invention by section 25(2) of the Patents Act.[29] However, this restriction applies "only to the extent to which a patent or an application for a patent relates to that thing as such"[30] and should not prevent, for example, a product, process, or method which may be implemented on a computer from being an invention, provided that the requirements of novelty and inventiveness are met.

South Korea

[edit]

In South Korea, software is considered patentable and many patents directed towards "computer programs" have been issued.[31] In 2006, Microsoft's sales of its "Office" suite were jeopardized due to a possible patent infringement.[32] A ruling by the Supreme Court of Korea found that patents directed towards automatic language translation within software programs were valid and possibly violated by its software.[32]

Thailand

[edit]

As like as 52(2) of the European Patent Convention (EPC), section 9 of the Thai Patent Act 1999 states that Thai patent law does not include software (or computer program) from patentability because the computer software is not considered as an “invention”, in which it is not the idea of the product itself. Hence, the software is considered as the manual or instruction that was controlled by users to perform the tasks.[33]

A software patents law in Thailand has been controversial debates among the economists and national developers’ overtime since there were two significant developments in the international patent law; (1) the European Union's attempt to harmonize national patent laws by the Proposal for a Directive of the European Parliament and Council on the patentability of computer-implemented inventions,[34] and (2) the US court decision to expand patent protection to business methods.[34] The opinions are divided into two sides. Dr. Tangkitvanich, the IT specialist of Thailand Development Research Institute (TDRI), raised his concern that Thailand is not in a good stage for a software patent as there were several flaws in patent rights. For example, the business method prevention has high tendency to hinder the growth in innovations especially for the infant software companies.[35] Moreover, the software patent may cause monopoly and innovation problems. “Monopoly will thwart innovations of new software products, particularly open-source software”, said by a group of Thai Economists. However, Dr. Hirapruk who is the Director of Software Park Thailand, on the other hand, provides his support on allowing the computer programs to be patentable: “Thailand had to provide a patent-right protection for computer software to ensure foreign high-tech investors that software producers' creativity would be secured from violations in Thailand”. As a result, Mr. Sribhibhadh, president of the Association of Thai Software Industry, emphasized that there will need to be a clear overview of the impact on the local industry if Thailand really had to fully implement the patent right protections.

United States

[edit]
Growth of software patents in US
In 2024, AI patents in China and the US numbered more than three-fourths of AI patents worldwide.[36] Though China had more AI patents, the US had 35% more patents per AI patent-applicant company than China.[36]

The first software patent was issued June 19, 1968 to Martin Goetz for a data sorting algorithm.[37] The United States Patent and Trademark Office has granted patents that may be referred to as software patents since at least the early 1970s.[38] In Gottschalk v. Benson (1972), the United States Supreme Court ruled that a patent for a process should not be allowed if it would "wholly pre-empt the mathematical formula and in practical effect would be a patent on the algorithm itself", adding that "it is said that the decision precludes a patent for any program servicing a computer. We do not so hold."[39] In 1981, the Supreme Court stated that "a claim drawn to subject matter otherwise statutory does not become nonstatutory simply because it uses a mathematical formula, computer program, or digital computer" and a claim is patentable if it contains "a mathematical formula [and] implements or applies the formula in a structure or process which, when considered as a whole, is performing a function which the patent laws were designed to protect".[40] When a patent application is examined by the USPTO, the initial threshold question (for each claim) is whether the subject matter is eligible, so this is evaluated separately and prior to the other patentability criteria (novelty, nonobviousness).[41] This is notably different than the European approach (see above).

Due to different treatment of federal patent rights in different parts of the country, in 1982 the U.S. Congress created a new court (the Federal Circuit) to hear patent cases. Following several landmark decisions by this court, by the early 1990s the patentability of software was well established, and in 1996 the USPTO issued Final Computer Related Examination Guidelines stating that "A practical application of a computer-related invention is statutory subject matter. This requirement can be discerned from the variously phrased prohibitions against the patenting of abstract ideas, laws of nature or natural phenomena" (emphasis added).[42]

The emergence of the Internet and e-commerce led to many patents being applied for and being granted for business methods implemented in software and the question of whether business methods are statutory subject matter is a separate issue from the question of whether software is. Critics of the Federal Circuit believe that the non-obviousness standard is partly responsible for the large increase in patents for software and business methods.[43] There have been several successful enforcement trials in the United States, some of which are listed in the list of software patents article.

An issue with software patent intellectual property rights is typically revolved around deciding whether the company or inventor owns it.

As a matter of law, in the United States, the employee generally owns the IP right unless the employee's inventing skills or task to create the invention is the main specific hiring reason or a specific clause in the employment agreement assigning invention rights.[44][45]

A work for hire created after 1978 has copyright protection for 120 years from its creation date or 90 years from its publication date whichever comes first.[46] Patent protection for software lasts 20 years.[47]

Indonesia

[edit]

In Indonesia, software cannot be protected by patents, until the implementation of the Law No. 13 Year 2016, Patent Law in Indonesia.[48] To begin evaluation, it is necessary to distinguish whether or not the application is considered an invention. Under Law No. 14 Year 2001, Article 1 of Patent Law in Indonesia,[49] application is considered as an invention if the activity is created to solve a particular conflict or problem in the technology sector. Furthermore, it can be executed in the medium of a new process or product or a developmental enhancement in a product or process. According to Law No. 14 Year 2001, Article 7 of Patent Law in Indonesia.,[49] an application can not be patented as an invention if the product or process contradicts or challenges the current regulations and rules, public order or ethics, and religious morality. In addition, if the application is treated as a method or theory in the scientific or mathematics, argued to be any type of living creatures, with the exception of micro-organisms, or is considered as an essential biological measure to produce plants or animals, the application is not a patentable invention.[49]

As software contains algorithms, it is deemed to be part of the field of mathematics; hence, software cannot be protected by patents in Indonesia. However, one way for the Indonesian Intellectual Property office to grant software patents in Indonesia is if the application has been patented in other nations, which have ratified the Patent Corporation Treaty (PCT). Therefore, in accordance to the regulations under the Patent Cooperation Treaty, a software will have a regional protection among the participating entities of World Intellectual Property Organization (WIPO).[50]

An important update was enacted on 26 August 2016, the Law No. 13 Year 2016, Patent Law in Indonesia.[48] This update is geared to encourage innovation and growth by augmenting the number of patents within the public and private sector in Indonesia. This update proposes an extension of protection for simple patent, which grants application for patents for new improvements or inventions to existing processes. Intangible inventions can also be patented; under the former law, simple patent is restricted for tangible inventions, which has a positive implication for software patents in Indonesia. Furthermore, these changes provide more protection to the pharmaceutical industry and encourage public access to medical knowledge. This can boost new software ideas and processes within the healthcare and pharmaceutical sector. This update provides a stronger protection of traditional knowledge. In addition, a significant update is the usability of electronic filling and electronic media. Under this new law, application can be made electronically.

Purpose of patents

[edit]

For the U.S., the purpose of patents is laid down in the constitutional clause that gives Congress the power to "promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries;" (Article I, Section 8, Clause 8).[51] For Europe, there is no similar definition. Commonly four patent justification theories are recognised, as laid down for instance by Machlup in 1958,[52] which include justice to the inventor and benefit for society by rewarding inventors. Disclosure is required in return for the exclusive right, and disclosure may promote further development. However, the value of disclosure should not be overestimated: some inventions could not be kept secret otherwise, and patents also prohibit independent reinventions to be exploited.

There is debate as to whether or not these aims are achieved with software patents.

Proposals

[edit]

In seeking to find a balance, different countries have different policies as to where the boundary between patentable and non-patentable software should lie. In Europe, a number of different proposals for setting a boundary line were put forward during the debate concerning the proposed Directive on the patentability of computer-implemented inventions, none of which were found acceptable by the various parties to the debate. Two particular suggestions for a hurdle that software must pass to be patentable include:

  • A computer program that utilises "controllable forces of nature to achieve predictable results".[53]
  • A computer program which provides a "technical effect".[54]

In the US, Ben Klemens, a Guest Scholar at the Brookings Institution, proposed that patents should be granted only to inventions that include a physical component that is by itself nonobvious.[55] This is based on Justice William Rehnquist's ruling in the U.S. Supreme Court case of Diamond v. Diehr that stated that "... insignificant postsolution activity will not transform an unpatentable principle into a patentable process."[56] By this rule, one would consider software loaded onto a stock PC to be an abstract algorithm with obvious postsolution activity, while a new circuit design implementing the logic would likely be a nonobvious physical device. Upholding an "insignificant postsolution activity" rule as per Justice Rehnquist's ruling would also eliminate most business method patents.

Obviousness

[edit]

A common objection to software patents is that they relate to trivial inventions.[57] A patent on an invention that many people would easily develop independently of one another should not, it is argued, be granted since this impedes development.

Different countries have different ways of dealing with the question of inventive step and non-obviousness in relation to software patents. Europe uses an 'Inventive step test'; see the Inventive step requirement in Europe and, for instance, T 258/03.

Criticism

[edit]

Compatibility

[edit]

There are a number of high-profile examples where the patenting of a data exchange standards forced another programming group to introduce an alternative format. For instance, the Portable Network Graphics (PNG) format was largely introduced to avoid the Graphics Interchange Format (GIF) patent problems, and Ogg Vorbis to avoid MP3. If it is discovered that these new suggested formats are themselves covered by existing patents, the final result may be a large number of incompatible formats. Creating such formats and supporting them costs money and creates inconvenience to users.

Computer-implemented invention (CII)

[edit]

Under the European Patent Convention (EPC), and in particular its Article 52,[58] "programs for computers" are not regarded as inventions for the purpose of granting European patents,[59] but this exclusion from patentability only applies to the extent to which a European patent application or European patent relates to a computer program as such.[60] As a result of this partial exclusion, and despite the fact that the EPO subjects patent applications in this field to a much stricter scrutiny[61] when compared to their American counterpart, that does not mean that all inventions including some software are de jure not patentable.

[edit]

Patent and copyright protection constitute two different means of legal protection which may cover the same subject matter, such as computer programs, since each of these two means of protection serves its own purpose.[62] Software is protected as works of literature under the Berne Convention. This allows the creator to prevent another entity from copying the program and there is generally no need to register code in order for it to be copyrighted.

Patents, on the other hand, give their owners the right to prevent others from using the technology defined by the patent claims, even if the technology was independently developed and there was no copying of a software or software code involved. In fact, one of the most recent EPO decisions[63] clarifies the distinction, stating that software is patentable, because it is basically only a technical method executed on a computer, which is to be distinguished from the program itself for executing the method, the program being merely an expression of the method, and thus being copyrighted.

Patents cover the underlying methodologies embodied in a given piece of software, or the function that the software is intended to serve, independent of the particular language or code that the software is written in. Copyright prevents the direct copying of some or all of a particular version of a given piece of software, but does not prevent other authors from writing their own embodiments of the underlying methodologies. Assuming a dataset meets certain criteria, copyright can also be used to prevent a given set of data from being copied while still allowing the author to keep the contents of said set of data a trade secret.[64]

Whether and how the numerus clausus principle shall apply to the legal hybrid software[65] to provide a judicious balance between property rights of the title holders and freedom rights of computing professionals[66] and society as a whole,[67] is in dispute.[68][69][70]

Debate

[edit]

There is a debate over the extent to which software patents should be granted, if at all. Important issues concerning software patents include:

  • Whether software patents should be allowed, and if so, where the boundary between patentable and non-patentable software should lie;[71]
  • Whether the inventive step and non-obviousness requirement is applied too loosely to software;[72] and
  • Whether patents covering software discourage, rather than encourage, innovation;[73]
  • Whether software based on mathematical methods may be allowed if the mathematics or algorithm in question is complicated enough and may not be implemented with pencil and paper.[74]

Open source software

[edit]

There is strong dislike in the free software community towards software patents. Much of this has been caused by free software or open source projects terminating[75] when the owners of patents covering aspects of a project demanded license fees that the project could not pay, or was not willing to pay, or offered licenses with terms that the project was unwilling to accept, or could not accept, because it conflicted with the free software license in use.[76]

Several patent holders have offered royalty-free patent licenses for a very small portion of their patent portfolios. Such actions have provoked only minor reaction from the free and open source software communities for reasons such as fear of the patent holder changing their mind or the license terms being so narrow as to have little use.[77] Companies that have done this include Apple,[78] IBM,[79] Microsoft,[80] Nokia,[81] Novell,[82] Red Hat,[83] and Sun (now Oracle).[84]

In 2005, Sun Microsystems announced that they were making a portfolio of 1,600 patents available through a patent license called Common Development and Distribution License.[85]

In 2006, Microsoft's pledge not to sue Novell Linux customers, openSUSE contributors, and free/open source software developers over patents[86] and the associated collaboration agreement with Novell[87] was met with disdain from the Software Freedom Law Center[88] while commentators from the Free Software Foundation stated that the agreement would not comply with GPLv3. Meanwhile, Microsoft has reached similar agreements with Dell and Samsung,[89] due to alleged patent infringements of the Linux operating system. Microsoft has also derived revenue from Android by making such agreements-not-to-sue with Android vendors.[90]

Unisys case

[edit]

In the late 1990s, Unisys claimed to have granted royalty free licenses to hundreds of not-for-profit organizations that used the patented LZW compression method and, by extension, the GIF image format. However, this did not include most software developers and Unisys were "barraged" by negative and "sometimes obscene" emails from software developers.[91]

Licensing

[edit]
Total US software patent counts by class of invention as of 2015[92]
US class Description Total patents issued
700 Data Processing: Generic Control Systems or Specific Applications 26042
701 Data Processing: Vehicles, Navigation, and Relative Location 38566
702 Data Processing: Measuring, Calibrating, or Testing 27130
703 Data Processing: Structural Design, Modeling, Simulation, and Emulation 10126
704 Data Processing: Speech Signal Processing, Linguistics, Language Translation, and Audio Compression/Decompression 17944
705 Data Processing: Financial, Business Practice, Management, or Cost/Price Determination 38284
706 Data Processing: Artificial Intelligence 9161
707 Data Processing: Database and File Management or Data Structures 47593
708 Electrical Computers: Arithmetic Processing and Calculating 9993
709 Electrical Computers and Digital Processing Systems: Multicomputer Data Transferring 56001
710 Electrical Computers and Digital Data Processing Systems: Input/Output 23991
711 Electrical Computers and Digital Processing Systems: Memory 34025
712 Electrical Computers and Digital Processing Systems: Processing Architectures and Instruction Processing (e.g., Processors) 10461
713 Electrical Computers and Digital Processing Systems: Support 30695
714 Error Detection/Correction and Fault Detection/Recovery 38532
715 Data Processing: Presentation Processing of Document, Operator Interface Processing, and Screen Saver Display Processing 25413
716 Computer-Aided Design and Analysis of Circuits and Semiconductor Masks 13809
717 Data Processing: Software Development, Installation, and Management 17336
718 Electrical Computers and Digital Processing Systems: Virtual Machine Task or Process Management or Task Management/Control 7615
719 Electrical Computers and Digital Processing Systems: Interprogram Communication or Interprocess Communication (Ipc) 5456
720 Dynamic Optical Information Storage or Retrieval 3877
725 Interactive Video Distribution Systems 12076
726 Information Security 21144
Total 525270

Patenting software is widespread in the US. As of 2015, approximately 500,000 patents had issued in the 23 classes of patents covering "computer implemented inventions" (see table).

Many software companies cross license their patents to each other. These agreements allow each party to practice the other party's patented inventions without the threat of being sued for patent infringement. Microsoft, for example, has agreements with IBM, Sun (now Oracle), SAP, Hewlett-Packard, Siemens, Cisco, Autodesk,[93] and recently Novell. Microsoft cross-licensed its patents with Sun, despite being direct competitors, and with Autodesk even though Autodesk has far fewer patents than Microsoft.

The ability to negotiate cross licensing agreements is a major reason that many software companies, including those providing open source software, file patents. As of June 2006, for example, Red Hat had developed a portfolio of 10 issued US patents, 1 issued European patent, 163 pending US patent applications, and 33 pending international PCT (Patent Cooperation Treaty) patent applications. Red Hat uses this portfolio to cross license with proprietary software companies so that they can preserve their freedom to operate.[83]

Other patent holders are in the business of inventing new "computer implemented inventions" and then commercializing the inventions by licensing the patents to other companies that manufacture the inventions. Walker Digital, for example, has generated a large patent portfolio from its research efforts, including the basic patent on the Priceline.com reverse auction technology. US universities also fall into this class of patent owners. They collectively generate about $1.4 billion per year through licensing the inventions they develop to both established and start up companies in all fields of technology, including software.[94]

Still other patent holders focus on obtaining patents from original inventors and licensing them to companies that have introduced commercial products into the marketplace after the patents were filed. Some of these patent holders, such as Intellectual Ventures, are privately held companies financed by large corporations such as Apple, Microsoft, Intel, Google, etc. Others, such as Acacia Technologies, are publicly traded companies with institutional investors being the primary shareholders.[95]

The practice of acquiring patents merely to license them is controversial in the software industry. Companies that have this business model are pejoratively referred to as patent trolls. It is an integral part of the business model that patent licensing companies sue infringers that do not take a license. Furthermore, they may take advantage of the fact that many companies will pay a modest license fee (e.g. $100,000 to $1,000,000) for rights to a patent of questionable validity, rather than pay the high legal fees ($2,000,000 or more) to demonstrate in court that the patent is invalid.[citation needed]

See also

[edit]

References

[edit]
[edit]
Revisions and contributorsEdit on WikipediaRead on Wikipedia
from Grokipedia
A software patent is a form of protection granted to inventors for novel, non-obvious software-related inventions, such as algorithms, computer-implemented methods, or systems that perform specific functions tied to technological improvements, typically under utility patent laws in jurisdictions like the where they must demonstrate practical application beyond mere abstract ideas. In the US, the first such patent was issued in 1968 to Martin Goetz for a data sorting program, marking the onset of patenting software innovations amid growing computer industry needs for exclusivity to recoup development costs. Landmark cases, including Diamond v. Diehr (1981) which upheld patents for software-integrated and Alice Corp. v. CLS Bank (2014) which invalidated claims for generic computer implementation of abstract concepts without inventive concepts, have shaped eligibility by requiring software patents to improve computer functionality or solve technical problems rather than merely automate mental processes. The practice remains highly controversial, with critics arguing that low-quality, overly broad software patents foster non-practicing entities ("patent trolls") that extract rents through litigation rather than , imposing costs estimated in billions annually on the tech sector and potentially hindering cumulative in fast-evolving fields like open-source programming. Empirical studies yield mixed results on net impact: some find positive correlations between software patenting and R&D in firms, suggesting incentives for disclosure and , while others indicate negligible or negative effects on overall rates, particularly in industries reliant on rapid where patents may deter or enable strategic blocking. Proponents counter that without patents, free-riding on disclosed inventions would undermine incentives for high-risk software R&D, as evidenced by historical tech firm reliance on patents for venture and market entry in the 1970s-1990s software boom. Internationally, approaches diverge— restricts pure software patents under the , emphasizing technical effects, while countries like largely exclude them to promote software —highlighting ongoing debates over balancing monopoly with in a sector where ideas propagate digitally at low .

Definition and Scope

Core Definition and Eligibility Criteria

A software patent grants exclusive rights to an where software implements a method, , , or apparatus that solves a technical problem, rather than protecting the software itself, which falls under law. Such patents typically claim functional aspects like algorithms or techniques integrated into hardware or computational environments. Eligibility for software patents requires compliance with fundamental standards: novelty, inventive step (non-obviousness), and industrial applicability (), alongside statutory subject matter requirements excluding judicial exceptions. , under 35 U.S.C. § 101, claims must qualify as a , , manufacture, or and pass the Alice/Mayo two-step framework: first, determining if the claim is directed to an abstract idea (e.g., mathematical concepts, organizing human activity, or mental ); second, evaluating whether additional elements integrate the idea into a practical application—such as improving computer functionality—or provide an inventive concept beyond routine computer implementation. For instance, claims reciting generic computers performing abstract mathematical operations without technical improvement are ineligible, whereas those demonstrating specific enhancements like self-referential structures for faster searches may qualify. In jurisdictions like the , eligibility hinges on a two-hurdle approach: the must exhibit technical character (e.g., involving hardware and aiming for a technical effect) and demonstrate inventive step in technical features only, excluding "computer programs as such" unless they contribute technically, such as optimizing circuit simulations via software models. Purely mathematical methods or business schemes implemented in software without technical contribution remain unpatentable. These criteria ensure software patents protect genuine technological advances rather than abstract or non-technical concepts, though application varies, with post-2014 U.S. scrutiny under Alice rejecting many claims lacking demonstrable technical integration. Software patents differ from protection primarily in scope and purpose. safeguards the specific expression of software code as a literary work, preventing unauthorized copying of the literal code or closely derived non-literal elements such as program structure, but it does not extend to the underlying ideas, algorithms, or functional processes implemented by the code. For software applications, particularly mobile or software apps, utility patents—equivalent to software patents—cover functional aspects such as processes, methods, or systems, while design patents protect ornamental user interface elements like icons or layouts. Copyright automatically protects code expression, and trademarks safeguard app names or logos, but utility patents provide stronger exclusivity by preventing equivalent functional implementations, even if developed independently with different code. In contrast, software patents protect the inventive functionality or technical method enabled by the software, such as a novel process for compression or , provided it meets criteria like novelty, non-obviousness, and industrial applicability; this allows patentees to exclude others from using equivalent implementations, even if developed independently with different code. Unlike trade secrets, which protect confidential software elements—such as proprietary algorithms or optimization techniques—indefinitely as long as reasonable secrecy measures are maintained, software patents require full public disclosure of the in exchange for a limited monopoly term of typically 20 years from filing. Trade secret protection offers no remedy against independent , , or accidental disclosure, making it suitable for software components that are difficult to discern from external use, whereas patents provide enforceable exclusivity against all infringers, including those who arrive at the same separately, but necessitate examination and potential invalidation challenges. Software patents are also distinguished from business method patents, which often involve abstract economic or organizational schemes; while both may employ software implementation, eligibility for software patents hinges on claiming a technical contribution or improvement, such as enhancing computer functionality, rather than merely automating a pre-existing business practice without inventive integration. , following the Supreme Court's decision in Alice Corp. v. CLS Bank, claims directed to abstract ideas—including many business methods—must demonstrate an "inventive concept" by improving technology or transforming the claim into a practical application to avoid ineligibility under 35 U.S.C. § 101, whereas pure software inventions solving technical problems (e.g., via specific data structures or processing efficiencies) face less stringent scrutiny if they exceed mere abstraction.

Technical and Inventive Requirements

In jurisdictions permitting software patents, inventions must demonstrate both technical character—establishing eligibility beyond mere abstract ideas or excluded subject matter—and an inventive step, ensuring non-obviousness relative to . Technical requirements typically demand integration into a practical application, such as improving computer functionality or solving a specific technical problem, rather than claiming algorithms or business methods in isolation. In the United States, under 35 U.S.C. § 101, software-related claims must fall within statutory categories (process, machine, manufacture, or composition of matter) and avoid judicial exceptions like abstract ideas. The two-step Alice Corp. v. CLS Bank framework applies: first, determine if the claim is directed to an abstract idea (e.g., fundamental economic practices or mental processes); second, if so, assess whether it includes additional elements amounting to significantly more, such as a technological improvement in computer capabilities or data processing efficiency. USPTO guidance updated as of August 4, 2025, emphasizes that claims reciting concrete improvements—like enhanced machine learning model training or specific hardware integrations—may qualify as eligible, while purely generic computer implementations of abstract concepts do not; examiners are instructed to limit rejections based on overly broad "mental process" analogies for AI/software innovations. For inventive step, U.S. law requires non-obviousness under 35 U.S.C. § 103, meaning the invention must not have been obvious to a person of ordinary skill in the art at the time of filing, considering the scope and content of , differences from it, and secondary factors like commercial success. Software claims often succeed by demonstrating unexpected technical results, such as reduced or novel data transformation methods, rather than routine automation of known processes. At the (EPO), computer programs "as such" are excluded from patentability per Article 52(2)(c) EPC, but claims gain eligibility through a "further technical effect" beyond the computer's normal operations, such as controlling an industrial process or optimizing resource allocation in a network. The COMVIK approach (T 0641/00) evaluates inventive step by isolating technical features from non-technical contributions (e.g., mathematical methods), requiring the technical implementation to involve a non-obvious solution to a technical problem, assessed via the problem-solution framework: identifying the closest , formulating the objective technical problem, and verifying if the solution would be obvious. Examples of patentable effects include enhancements or systems, whereas pure data presentation or rule-based simulations typically fail unless tied to verifiable technical contributions. Across jurisdictions, applicants must provide evidence of technical implementation details in specifications, such as flowcharts or demonstrating how software interacts with hardware or achieves measurable efficiencies, to substantiate both eligibility and inventiveness. Failure to articulate these often leads to rejections, as seen in EPO Board of Appeal decisions emphasizing that inventive merit resides in the concrete technical realization, not the underlying idea.

Historical Development

Origins and Early Precedents

The concept of patenting inventions involving software emerged alongside the development of electronic computers in the mid-20th century. In the late and early , patent applications for devices often included claims encompassing programmed operations or algorithms as integral components of hardware processes, treating software as a functional element of the rather than an abstract idea separable from physical implementation. These early efforts reflected the era's understanding of computers as specialized machinery, where control mechanisms—effectively rudimentary software—were patented within broader apparatus claims to satisfy statutory requirements for tangible, useful inventions. By the 1960s, as commercial software markets developed and programs became commoditized products separate from hardware, patent seekers adapted strategies such as "embodying" software in claims tied to systems or methods to navigate and Office (PTO) examiner skepticism toward purely mathematical or mental processes. The and Office (USPTO) granted its first explicit software patent, U.S. Patent No. 3,380,029, on April 30, 1968, to Martin Goetz of Applied Data Research, Inc., for a method and system for sorting data records using a computer, marking a for protecting algorithmic processes implemented via programming. This issuance occurred despite internal USPTO guidelines issued earlier that year cautioning against patenting computer programs per se, as they risked encompassing non-statutory abstract ideas under 35 U.S.C. § 101. Lower court decisions in the late further established early precedents supporting when tied to practical applications. In Prater & Wei v. Thiessen Metals Corp. (1968), a federal district court upheld the validity of claims for a computer-implemented process for analyzing metallurgical data, affirming that programmed computations could constitute if they produced concrete, transformative results rather than mere calculations. Such rulings encouraged applicants to frame software innovations as improvements in technical fields like or , laying groundwork for broader acceptance amid growing industry reliance on proprietary code for . However, these precedents coexisted with persistent rejections, highlighting ongoing debates over whether software's intangible nature aligned with traditional patent eligibility for processes and machines.

Pivotal US Supreme Court Cases

In Gottschalk v. Benson (1972), the unanimously held that an algorithm for converting numbers into pure binary numbers was not patent-eligible subject matter under 35 U.S.C. § 101, as it constituted an abstract mathematical formula that risked preempting all uses of the formula in . The invention involved a method programmable on a general-purpose digital computer for any use, leading the Court to reject patentability to avoid monopolizing the algorithm itself. The Court revisited algorithmic patentability in Parker v. Flook (1978), ruling 5-3 that a method for updating maximum and minimum alarm limits in a catalytic chemical conversion process—novel only due to a mathematical formula for calculating the limits—was ineligible for patent protection. Unlike prior art processes, the claim's innovation lay solely in post-solution activity of applying the formula, which the majority deemed insufficient to transform an abstract mathematical algorithm into statutory subject matter. This decision emphasized that the novelty of a claimed process cannot derive exclusively from a non-patentable mathematical formula. Diamond v. Diehr (1981) marked a partial shift, with the holding 5-4 that a process for curing using a computer-programmed to monitor and adjust curing time was patent-eligible. The invention integrated the mathematical formula into a physical process transforming raw rubber into cured products, distinguishing it from pure algorithms by claiming an industrial process rather than the equation in isolation. The majority clarified that statutory processes under § 101 include those producing a "new and useful" result when tied to tangible transformation, even if involving software-controlled steps. In Bilski v. Kappos (2010), the Court unanimously affirmed the ineligibility of a method for hedging risk in commodities trading as an abstract idea, rejecting the Federal Circuit's machine-or-transformation test as the exclusive § 101 criterion but upholding its utility as a significant clue to patentability. The 5-4 majority opinion, joined by a , warned against overbroad method patents that could "pre-empt" fundamental economic practices, influencing software claims by reinforcing scrutiny of abstract ideas without narrowing eligibility for technological inventions. The landmark Alice Corp. v. CLS Bank International (2014) decision, unanimous in outcome, invalidated patents on a computer-implemented to mitigate in financial transactions, establishing a two-step framework for § 101 analysis: first, determine if the claim is directed to a patent-ineligible abstract idea; second, examine if it includes an "inventive concept" transforming the idea into a patent-eligible application. The Court found Alice's claims merely implemented an abstract idea of intermediated settlement on generic computers without adding significantly more, as routine elements failed to confer eligibility. This framework has since invalidated numerous software-related patents, prompting the USPTO to issue guidance emphasizing that generic computer implementation does not overcome abstract idea ineligibility.

Global Historical Milestones

The Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS), effective from 1995, established minimum standards for intellectual property protection among World Trade Organization members, requiring patents for inventions in "all fields of technology" under Article 27 without discrimination as to place of invention or technology, though it explicitly permitted exclusions for computer programs if deemed necessary to protect public order or morality, leaving software patentability to national laws rather than mandating uniform global protection. In Europe, the 1973 European Patent Convention (EPC), which entered into force in 1977, marked an early exclusionary stance by stipulating in Article 52(2)(c) that "programs for computers" are not regarded as inventions and thus unpatentable "as such," a provision intended to bar pure software claims while allowing broader inventive concepts; however, from 1985, the (EPO) shifted toward granting patents for computer-implemented inventions exhibiting a "technical effect" or solving a technical problem, as clarified in decisions like T 208/84 (), influencing over 30,000 such grants by the early despite ongoing debates. A 2002 proposal for a directive to harmonize of computer-implemented inventions across member states, which sought to codify technical contribution requirements, failed after the rejected it in 2005 by a vote of 409 to 149, citing risks of over-patentization and preserving the EPO's case-by-case . Japan pioneered permissive frameworks in , with the Patent Office issuing 1978 Examination Guidelines that enabled patents for software-related inventions contributing to technical fields or hardware improvements, building on earlier practices; this evolved through 1993 guideline revisions emphasizing "creation of technical ideas utilizing natural laws" for software, culminating in 2002 Patent Act amendments that explicitly affirmed computer programs as patentable products when industrially applicable, leading to a surge in filings from 1,200 software-related patents in to over 20,000 annually by the . In , the State Office's 2006 Guidelines for Examination represented a milestone, permitting patents for computer programs that provide "technical solutions" to technical problems, typically via integration with hardware, reversing prior non-grant practices and aligning with TRIPS obligations, which spurred growth to approximately 100,000 annual software-related applications by 2020. (Note: While is not cited as primary, corroborated by policy analyses; primary guidelines verifiable via SIPO archives.) India adopted an exclusionary approach under its 2005 Patents (Amendment) Act, inserting Section 3(k) to bar patents for "a mathematical or business method or a computer programme per se or algorithms," reflecting concerns over stifling innovation in its software services sector, with courts upholding this in cases like Ferid Allani v. Union of India (2019), prioritizing for pure software over patents.
YearMilestoneJurisdiction/BodyKey Impact
1973EPC signed, excluding computer programs "as such"Set foundational bar on pure software, enabling technical-effect exceptions.
1978JPO Guidelines permit software in technical inventionsFacilitated early Asian patenting of software-hardware combinations.
1995TRIPS Agreement effectiveGlobal (WTO)Established non-discrimination principle without requiring software patents.
2002Japan Patent Act amendment affirms programs as inventionsExpanded to standalone software with industrial utility.
2005EU Software Patent Directive rejected; India excludes "per se" programsEU/IndiaPreserved fragmented ; reinforced India's copyright focus.
2006China SIPO Guidelines allow technical software solutionsOpened patent pathway for integrated inventions.

Jurisdictional Frameworks

In the , patent eligibility for software-related inventions is primarily governed by 35 U.S.C. § 101, which excludes abstract ideas, laws of nature, and natural phenomena from protection, requiring inventions to be processes, machines, manufactures, or compositions of matter. Software claims directed to mere automation of mental processes or mathematical algorithms on generic computers are ineligible, but those demonstrating a technological improvement or practical application may qualify. The U.S. and Office (USPTO) examines such applications under the same standards as other technologies, emphasizing novelty, non-obviousness, and enablement alongside subject matter eligibility. Early precedents established boundaries: In Gottschalk v. Benson (1972), the invalidated claims for a pure converting numerals to binary, deeming it an unpatentable mental process regardless of computer implementation. Similarly, Parker v. Flook (1978) rejected a formula updating alarm limits in a catalytic conversion process, holding that adding insignificant post-solution activity did not confer eligibility. However, Diamond v. Diehr (1981) upheld a rubber-curing process using a computer to repeatedly calculate cure time via the , as it transformed raw rubber into a finished product, integrating the formula into an industrial process. The Federal Circuit expanded eligibility in the . In re Alappat (1994) treated software-defined hardware as a statutory , not mere manipulation. State Street Bank & Trust Co. v. Signature Financial Group (1998) validated a system for financial portfolio , rejecting categorical exclusion of business methods and affirming that practical, non-abstract applications are patentable. This facilitated a surge in software and business method patents during the late and early . Subsequent Supreme Court rulings refined the framework. Bilski v. Kappos (2010) rejected the Federal Circuit's rigid "machine-or-transformation" test as the sole § 101 criterion but invalidated a method for hedging as an abstract idea lacking transformation or machine ties. The landmark Alice Corp. v. CLS Bank International (2014) established a two-step test: first, determine if claims are directed to a patent-ineligible abstract idea (e.g., fundamental economic practices or generic computer functions); second, assess whether additional elements provide an "inventive concept" transforming the idea into a patent-eligible application, such as by improving or solving a specific technical problem. Claims reciting software on generic hardware for routine tasks typically fail, leading to invalidation of thousands of patents in litigation. Post-Alice, USPTO guidance has evolved to clarify eligibility, with 2019 revisions emphasizing integration into practical applications and 2024 updates addressing AI inventions by requiring claims to improve technology rather than merely use computers as tools. Examiners apply the Alice framework during prosecution, often rejecting broad claims but allowing those tied to specific hardware improvements or novel . As of mid-2024, approximately 61% of issued U.S. patents contain software-related claims, reflecting steady grant rates despite heightened scrutiny. Total utility patent grants reached 324,042 in 2024, up 4% from 2023, with software innovations prominent in fields like AI and . Courts continue to apply Alice strictly, invalidating vague claims while upholding those with concrete technical contributions, fostering debate over whether this balances innovation incentives against over-patenting risks.

European Union and EPO

In the (EPC), administered by the (), Article 52(2)(c) excludes "programs for computers" from , alongside other non-inventions such as discoveries, mathematical methods, and business methods. However, Article 52(3) clarifies that this exclusion applies only to such programs "as such," meaning computer-implemented inventions (CII) remain eligible if they exhibit technical character beyond mere software implementation, such as contributing a technical effect that solves a technical problem in a non-obvious manner. The EPO assesses CII under the standard criteria of novelty, inventive step, and industrial applicability, requiring claims to demonstrate a concrete technical contribution, often tied to hardware improvements, efficiency, or control of physical processes rather than abstract algorithms or pure . EPO case law, developed by the Boards of Appeal, refines this approach; for instance, in decision T 0641/00 (Vicom), image processing software was deemed patentable due to its technical effect on data representation, establishing that non-physical technical effects can suffice if they alter technical processes. More recently, Enlarged Board decision G 1/19 (2021) held that computer simulations of technical systems or processes are examinable as CII, provided they produce verifiable technical effects akin to physical inventions, rejecting blanket exclusions for simulation methods. The EPO's Guidelines for Examination, updated in April 2025, emphasize claiming CII in terms of functional technical features, with examples including neural networks for technical tasks like in machinery, but excluding pure mathematical models without implementation. This framework has enabled thousands of CII grants annually, though rejection rates remain high for claims lacking discernible technicality, with appeals often hinging on COMVIK guidance (T 258/03) that bifurcates technical and non-technical elements for inventive step evaluation. In the European Union, patent granting falls under national laws or the EPC via the EPO, with no supranational EU patent office dictating software eligibility beyond EPC rules; the unitary patent system, effective from June 2023, applies EPO criteria uniformly across participating EU states. A 2002 EU Commission proposal for a Directive on the Patentability of Computer-Implemented Inventions sought to codify and harmonize national practices, aiming to clarify technical contribution requirements while preventing broad "software as such" patents, but it was rejected outright by the European Parliament on July 6, 2005, amid concerns over over-patentability stifling open-source innovation. Consequently, EU software protection defaults to copyright under Directive 2009/24/EC (codifying 1991/250/EEC), which safeguards program code and preparatory materials but explicitly precludes patent-like idea protection, leaving substantive patent policy to EPC jurisprudence. National courts in EPC states, such as Germany's Federal Patent Court, often align with EPO decisions but may diverge slightly on enforcement, contributing to perceived inconsistencies despite the centralized grant process. This rejection preserved a cautious stance, prioritizing technicality to avoid the expansive software patenting seen elsewhere, though critics argue it hampers EU competitiveness in fields like AI and fintech.

Asia-Pacific Economies

In , computer software has been explicitly patentable since a 2002 amendment to Article 2(3) of the Patent Act, which defines inventions as creations of technical ideas utilizing laws of nature, including programs for computers. The Patent Office examines software-related inventions under guidelines requiring a technical character, such as producing a specific technical effect or solving a technical problem, rather than mere abstract algorithms. This approach allows claims encompassing both method and product forms, including means-plus-function formats, facilitating broader protection compared to some jurisdictions. China's patent law permits software-related inventions if they demonstrate technical features that solve a technical problem using technical means to achieve a technical solution, as outlined in the State Intellectual Property Office (SIPO) guidelines. Pure rules, methods for intellectual activities, or mathematical models remain ineligible, but inventions integrating software with hardware or processes yielding tangible technical improvements—such as enhanced data processing efficiency—are grantable. This framework has supported a surge in filings, with consensus emerging that software should not be categorically excluded if tied to inventive technical contributions. South Korea allows patents for software inventions when embodied on a storage medium or combined with hardware to form an integrated representing a technical advancement, per Korean Intellectual Property Office (KIPO) practice. Standalone computer programs fall under protection rather than patents, but method claims broadened by a 2020 Patent Act amendment incentivize enforcement for software-implemented processes with industrial applicability. Examination emphasizes novelty and inventive step, aligning software eligibility with overall invention criteria without explicit exclusions for computational elements. Australia grants patents for computer-implemented inventions that go beyond mere schemes or business methods, requiring a specific technical contribution or artificial effect, as clarified in Federal Court decisions and IP Australia guidelines updated through 2025. Software must demonstrate integration with hardware or processes yielding non-obvious improvements, such as novel data manipulation in industrial contexts, to satisfy manner-of-manufacture requirements under the Patents Act 1990. Despite evolving , including rulings rejecting abstract implementations, eligible inventions often involve tangible enhancements like improved machine functionality. India excludes "a mathematical or business method or a computer programme per se or algorithms" from under Section 3(k) of the Patents Act, 1970, but permits protection for software-embedded inventions demonstrating technical advancement tied to hardware or novel processes. The requires disclosure of sufficient non-software elements contributing to industrial applicability, as affirmed in guidelines and court interpretations rejecting pure software claims. This restrictive stance persists, with over 50% of emerging technology patents in the past decade involving software hybrids, though eligibility hinges on proving inventive synergy beyond code alone. In , software qualifies as patentable if it embodies a creation of technical ideas utilizing natural laws, per Article 21 of the Patent Act, with examination guidelines emphasizing technical effects beyond standard computer operations. Revised in 2021, these guidelines favor mixed claims incorporating hardware or specific algorithmic applications producing measurable improvements, such as in AI processing. methods implemented via software remain ineligible unless demonstrating technical character, aligning Taiwan's approach with regional emphasis on substantive technological contributions. Across economies, software patent filings have risen, particularly in AI and digital technologies, with and leading global shares as of 2024 data on AI-related patents. This trend reflects policy adaptations to foster innovation in high-tech sectors, though lags due to varying technicality thresholds and exclusions for abstract inventions.

Other Key Jurisdictions

In , computer-implemented inventions, including software, are provided they are not directed solely to abstract ideas, mere schemes, or rules of while operating on a computer, but instead demonstrate a tangible improvement such as enhanced computer functionality or a physical effect. The Canadian Office (CIPO) examines claims for novelty, non-obviousness, and utility, requiring evidence of a "physical existence or manifestation" beyond generic computing, as established in precedents like Canada Ltd. v. Commissioner of Patents (1981) and refined in Benjamin Technologies Corp. v. Amazon.com Inc. (2023), where the Federal Court of Appeal invalidated claims lacking such elements. Recent review in 2025 continues to assess eligibility thresholds, emphasizing distinctions from unpatentable mathematical methods. In , Sections 3(c) and 3(k) of the Patents Act, 1970 explicitly exclude the mere discovery of a scientific principle or the formulation of an abstract theory, and "a mathematical or method or a computer programme per se or algorithms" from , reflecting policy concerns over monopolizing abstract ideas amid rapid technological evolution. However, software-integrated inventions qualify if they produce a demonstrable technical effect, such as improved hardware operation or a novel solution to a technical problem, rather than mere of mental acts; the applies this via pre-grant opposition and examination guidelines updated in 2017 and 2024. This framework has resulted in approvals for embedded systems but rejections for standalone apps, with over 1,000 software-related applications filed annually as of 2023, though grant rates remain low due to strict inventive step scrutiny. In Brazil, Article 24, I of Industrial Property Law (Law No. 9,279/1996) bars patents on "computer programs as such," prioritizing copyright registration for source code via the National Institute of Industrial Property (INPI), which processed over 10,000 software registrations in 2023. Patent eligibility extends to computer-implemented inventions exhibiting a concrete technical application, such as algorithmic control of physical processes or hardware innovations, provided claims avoid pure data processing without inventive contribution; INPI guidelines from 2021 emphasize distinguishing technical effects from excluded abstract schemes. This dual regime supports hybrid protections, with patent grants rising 15% for tech-integrated claims between 2020 and 2024 amid Brazil's TRIPS compliance.

Theoretical and Economic Rationale

Incentives for Innovation and Investment

Proponents of software patents contend that they incentivize by conferring temporary exclusive , enabling inventors to recover high fixed costs associated with (R&D) in an industry characterized by rapid and low replication expenses once code is disclosed. Without such protection, software innovations—often comprising abstract algorithms or processes—face immediate copying risks through or parallel development, leading to under due to the public goods problem where innovators cannot fully appropriate returns. This rationale aligns with economic theory positing patents as a mechanism to internalize externalities, fostering causal chains from idea generation to by aligning private incentives with social benefits of technological progress. In practice, software patents purportedly encourage R&D allocation toward novel, patentable inventions rather than incremental improvements, as exclusivity allows pricing above to fund future projects. For high-tech firms, where software underpins core value, patents mitigate hold-up risks from competitors, promoting sustained investment cycles; theoretical models using real options frameworks demonstrate how patent certainty enhances the of R&D pipelines by hedging against appropriation. Empirical analyses in peer-reviewed literature link software patent holdings to elevated firm-level metrics, such as patent citations and product launches, suggesting a positive causal influence on inventive activity in competitive markets. Regarding , software s function as credible signals of , reducing investor uncertainty and facilitating capital inflows in asset-light sectors reliant on intangible assets. Venture-backed software startups, for example, leverage portfolios to secure funding, with studies showing s correlate with higher valuation multiples and lower by serving as collateral in financing arrangements. Cross-sectional data from U.S. software firms indicate that stronger protections post-1980s reforms coincided with accelerated private R&D expenditures, reaching approximately 15-20% of revenues in patent-intensive subsectors by the early , underscoring s' role in bridging the gap between and market viability.

Property Rights and Market Dynamics

Software patents confer upon inventors a bundle of exclusionary rights akin to , prohibiting unauthorized parties from making, using, offering for sale, selling, or importing the patented within the , typically for 20 years from the filing date under 35 U.S.C. § 154(a)(1). This legal monopoly addresses the inherent public goods characteristics of software—low replication costs and ease of —by enabling rights holders to internalize returns on investments that often exceed millions in development expenses, as evidenced by over 25,000 annual U.S. software patent applications reported in early analyses. Unlike copyrights, which safeguard only expressive elements and permit inter-firm replication of underlying ideas, patents protect functional innovations, providing stronger deterrence against competitive and fostering incentives for high-risk R&D in idea-intensive fields. These property rights reshape market dynamics by creating tradable assets that underpin licensing ecosystems and strategic alliances. In high-tech industries, software patents serve as "currency" for transactions, with cross-licensing agreements—such as those between and encompassing software technologies—allowing firms to exchange access rights, avert infringement suits, and accelerate product without full-scale litigation. This facilitates market entry for resource-constrained startups, which leverage patent assignments or licenses to attract and mitigate risks, as patents signal credible technological barriers and enhance firm valuations; for example, Apple's post-2007 patent portfolio contributed to rapid market dominance by deterring direct copies and enabling defensive positioning against entrants like . Aggregate data underscore this: U.S. IP licensing revenues, heavily driven by software-related patents, reached $115.2 billion in 2012 across 28 industries, supporting IP-intensive sectors that generated 5.3% of GDP and 3.9 million jobs. Property rights in software patents also influence competitive landscapes by promoting circumvention over replication, as exclusivity compels rivals to around disclosed inventions per 35 U.S.C. § 112 requirements, yielding diversified offerings and reduced . While granting temporary pricing power to recoup fixed costs—potentially elevating short-term consumer prices—these rights counteract free-riding, where absent protection, innovators might withhold disclosures or abandon projects, stifling cumulative progress in networked markets like operating systems and applications. Empirical patterns in the 1990s-2010s, with U.S. issuing over 100,000 patents annually amid litigation rates below 2%, indicate that such dynamics sustain investment flows without broadly impeding entry, as licensing markets absorb only a fraction of disputes into productive exchanges.

Empirical Support for Pro-Patent Positions

Empirical analyses have identified associations between software patent holdings and enhanced firm performance in technology sectors. For instance, a study of high-tech firms found that software patents contribute to through licensing markets, with broader technological generality correlating to higher licensing propensity, thereby facilitating and revenue generation beyond mere defensive use. In the industry, which encompasses , empirical on listed companies demonstrated that invention s, including those for software-related innovations, positively influence revenue and profit margins, with a one-unit increase in patent count linked to measurable gains after controlling for firm size and R&D expenditure. Software patents serve as signals of quality to investors, particularly benefiting smaller firms. examining patent applications by software startups revealed that patents enable exclusionary against larger incumbents, contradicting claims of uniform hindrance and supporting sustained high R&D investment levels in the sector, which averaged over 15% of revenues for many firms in the sampled period. This exclusionary power aids in attracting ; data from startup patent portfolios indicate that patented software s correlate with improved fundraising outcomes and higher exit valuations, as patents reduce appropriation risks and enhance bargaining positions in mergers or acquisitions. In market valuation contexts, software patents exhibit positive effects on returns. An of strategic patent utilization showed that software portfolios elevate firm , with empirical models confirming a direct premium for firms employing patents offensively against rivals, independent of overall patent volume. Broader economic contributions from patent-intensive industries, including software, underscore this: in 2010, such sectors generated $763 billion in value added (5.3% of U.S. GDP) and supported 3.9 million jobs, with software patents playing a key role in high-tech licensing and . These findings align with evidence from datasets, where software patents actively enable innovation markets rather than solely serving litigation avoidance.

Criticisms and Debates

Claims of Innovation Hindrance

Critics of software patents contend that they impose significant barriers to in the software sector, where rapid iterative development and cumulative improvements are central to progress. Unlike physical inventions, software innovations often build incrementally on , leading to overlapping claims that create "patent thickets"—dense webs of intersecting patents that raise transaction costs for licensing and deter new entrants. A study examining patent thickets in technology sectors found that such overlaps particularly burden , complicating commercialization and increasing the risk of inadvertent infringement. Empirical analyses indicate that software patents correlate with reduced (R&D) intensity at the firm level, suggesting they may substitute for genuine efforts by enabling defensive or litigious strategies over substantive investment. James Bessen and Michael J. Meurer, in their 2008 book Patent Failure, analyzed U.S. data and litigation records, concluding that the abstract nature of software patents undermines their function as clear property rights, resulting in high enforcement costs that disproportionately affect smaller firms and non-manufacturing patent holders (often termed "patent trolls"). Their findings show that for software-intensive industries, the economic benefits of patents accrue mainly to large incumbents through litigation, while costs—averaging millions per suit—discourage startups from pursuing novel ideas. The U.S. Federal Trade Commission's 2003 report To Promote highlighted software-specific issues, such as low patent quality due to examination challenges for abstract ideas, fostering "hold-up" problems where patentees extract rents from implementers post-investment, thus chilling follow-on . For instance, broad software patents have been cited in cases like Apple's slide-to-unlock interface (U.S. Patent No. 8,046,721, granted 2011), which prompted litigation against competitors including , allegedly delaying market entry for similar user-friendly features despite their intuitive nature. Critics argue this exemplifies how software patents reward trivial or obvious implementations over inventive steps, with data from the onward showing a surge in such grants correlating with heightened infringement suits rather than proportional R&D growth. Further evidence from sector analyses posits that software's low marginal reproduction costs and models (e.g., via open-source licensing) thrive without patents, as proprietary claims fragment ecosystems and impede sharing. A 2004 empirical review of the U.S. software patent "experiment" post-1980s liberalization found no clear boost to metrics like productivity gains, but elevated litigation rates—peaking at over 4,000 software-related suits annually by the early —that diverted resources from development. These claims are bolstered by observations in fast-evolving fields like mobile apps, where overlapping algorithmic patents have reportedly stalled incremental advances, though proponents counter with mixed findings on .

Litigation and Enforcement Challenges

Litigation of software patents encounters significant hurdles primarily due to the abstract and functional nature of software inventions, which complicates delineation of claim scope and proof of infringement. In the United States, the Supreme Court's 2014 decision in Alice Corp. v. CLS Bank International established a framework under 35 U.S.C. § 101 that deems many software patents ineligible if they claim abstract ideas implemented on generic computers without an inventive concept, leading to frequent invalidations. Post-Alice, the Federal Circuit has invalidated software patents in the majority of reviewed cases, with early statistics showing invalidation upheld in all but one instance by late 2015, and by 2020, courts rejecting approximately 85% of software patent appeals. Enforcement is further impeded by high litigation costs and the prevalence of non-practicing entities (NPEs), often termed patent trolls, which disproportionately target the software sector. These entities assert broad or ambiguously worded against operating companies, extracting settlements to avoid protracted disputes; annual direct out-of-pocket costs to defendants reached $29 billion as of recent estimates, with additional wealth destruction exceeding $60 billion yearly. owners' success rates in U.S. litigation hovered around 32% in 2023, reflecting challenges in establishing validity amid searches complicated by the rapid evolution and open-source proliferation in . Proving direct infringement poses technical difficulties, as software often operates across distributed systems or involves third-party components, invoking doctrines like divided infringement where multiple actors contribute to the patented process. This requires demonstrating control or joint liability, which courts interpret stringently, increasing evidentiary burdens such as analysis and expert testimony on functionality equivalence. Internationally, harmonization gaps exacerbate enforcement, as software patent eligibility varies—e.g., stricter in the —leading to fragmented protection and higher cross-border assertion costs. Overall, these factors contribute to a landscape where software patent assertions frequently fail or settle uneconomically, deterring some innovators while prompting strategic drafting to emphasize technical improvements over mere automation. Despite this, well-drafted patents focusing on specific technological solutions have shown enforceability in select 2025 cases, underscoring the importance of overcoming Alice scrutiny through concrete implementations.

Alternative IP Mechanisms and Overlaps

Copyright law provides automatic protection for software as a literary work, safeguarding the specific expression of code rather than underlying ideas or functionality. In the United States, this protection arises upon fixation in a tangible medium, with registration offering evidentiary benefits and eligibility for statutory damages under the Act of 1976. Unlike patents, copyright does not require novelty examination or disclosure of methods, enabling rapid, low-cost defense against direct copying, as evidenced by successful litigations like Oracle v. , where aspects of structure were deemed protectable expression. However, it fails to prevent independent development of similar functionality, limiting its scope compared to patents' monopoly on inventions. Trade secrets offer perpetual protection for confidential software elements, such as algorithms or proprietary datasets, provided reasonable secrecy measures are maintained, as defined under the adopted in 48 U.S. states. This mechanism avoids public disclosure required by patents, preserving competitive edges indefinitely until breach or independent discovery, with enforcement through claims yielding remedies like injunctions and . For software, firms like have analogously protected formulas, and empirical analyses show trade secrets comprising a significant portion of IP value in tech sectors where is costly. Yet, they provide no defense against lawful or parallel invention, vulnerabilities highlighted in cases like Waymo v. , where self-driving software secrets were allegedly stolen. Contractual mechanisms, including nondisclosure agreements (NDAs) and end-user license agreements (EULAs), complement these by enforcing usage restrictions beyond statutory IP limits. NDAs bind parties to confidentiality, while EULAs govern distribution and prohibit , as upheld in ProCD v. Zeidenberg (1996), affirming shrink-wrap licenses' enforceability. Trademarks protect software branding, such as names or logos, under the , deterring consumer confusion but not core innovations. These tools enable layered protection, with surveys of software firms indicating heavy reliance on contracts for commercialization without patent disclosure risks. Open source licensing emerges as a deliberate alternative, relinquishing exclusive control via permissive (e.g., MIT) or (e.g., GPL) models to foster collaborative development. The GPL, for instance, requires derivative works to remain open, indirectly mitigating patent threats through viral sharing, as seen in Linux's dominance without broad patenting. Empirical data from the Software Industry Association shows powering 96% of top websites by 2023, suggesting efficacy in rapid iteration absent patent monopolies. Overlaps arise when patents cover -implemented inventions, prompting defensive patent grants in licenses like 2.0 to immunize contributors against infringement suits. These mechanisms overlap with in hybrid strategies: copyrights shield code while target abstract methods, as clarified in the U.S. and Office's guidelines post-Alice Corp. v. CLS (2014), allowing complementary filings. Trade secrets can safeguard non-patented implementations, but patent disclosure may forfeit secrecy, creating trade-offs analyzed in resource-based firm studies favoring for defensible, non-secret innovations. Empirical firm-level data indicates software correlate with higher valuation in venture-backed startups, implying alternatives alone may underprotect against rivals' independent inventions, though critics note copyrights suffice for expression-heavy software with lower litigation costs.

Rebuttals and Evidence-Based Defenses

Proponents of software patents counter assertions that they stifle by citing empirical analyses linking patent availability to elevated (R&D) activity. A study of U.S. firms during the , when legal shifts expanded software patent eligibility, found a residual increase in patent propensity attributable to improved cost-effectiveness of such protections, which in turn correlated with higher R&D expenditures per employee in software-intensive industries. This evidence aligns with incentive theory, positing that excludable rights encourage investment in intangible assets like algorithms and systems, where replication costs are low but development expenses high. Criticisms regarding insufficient incentives for upstream innovation are rebutted by firm-level data showing software patents attract external capital, particularly venture funding. Research analyzing patent filings and investment flows demonstrates that software patents serve as credible signals of proprietary advantage, positively influencing investor decisions and enabling commercialization of complex inventions that might otherwise face free-rider risks. For startups, empirical patterns indicate patents enhance fundraising probabilities and valuation multiples in software sectors, with effects strongest when filings precede funding rounds by 1-2 years. Litigation and "patent troll" concerns, while acknowledged, are addressed by evidence that software patents underpin success rather than merely enabling . Longitudinal data from the U.S. IT hardware industry (1996-2015) reveal a positive relationship between software-based portfolios and subsequent new product announcements, suggesting mechanisms facilitate market entry and competitive differentiation over obstruction. Broader reviews of affirm that software-specific outcomes do not systematically diverge from positive impacts observed across technologies, with private valuations of software patents often exceeding filing costs when tied to verifiable technical contributions. Reforms targeting abusive practices, such as post-grant reviews, mitigate downsides without negating core protective benefits. Defenses further emphasize that alternatives like trade secrecy or inadequately safeguard interconnected software ecosystems, where disclosure requirements enable cumulative progress. Post-2014 Alice Corp. v. CLS Bank scrutiny, which invalidated abstract ideas without inventive application, prompted a shift toward higher-quality filings; subsequent analyses show this refined eligibility boosted surviving s' economic value and spurred adaptive R&D strategies among patentees. These findings rebut blanket prohibitions by illustrating how targeted patent regimes foster verifiable inventive activity over unmoored criticism.

Empirical Assessments

Positive Impacts on R&D and Commercialization

Software patents grant inventors temporary exclusivity over novel algorithms, methods, and systems, enabling firms to recoup substantial R&D costs through . Empirical analysis demonstrates that the "patent premium"—the additional private returns from patenting—positively influences R&D intensity, particularly in sectors where protection is effective, as firms weigh the of exclusivity against disclosure costs. In industries, which heavily rely on software innovations, patents correlate with higher firm revenue and profit, suggesting that protection enhances the financial viability of R&D investments in . For startups, software patents serve as signals of technological quality, facilitating access to and improving growth metrics. Patented startups exhibit 55% higher employment growth and 80% higher sales growth over five years compared to non-patented peers, with patents providing securable assets that attract external financing by offering salvage value in failure scenarios. Moreover, patents reduce risk for new firms while increasing the likelihood of exits through mergers or acquisitions, thereby supporting sustained efforts. Commercialization benefits extend to licensing markets, where software patents enable efficient and further development. In high-tech sectors, these patents underpin active markets for rights trading, with 98–99% of issued patents deployed commercially rather than involved in litigation, contributing to in patent-intensive industries equivalent to $763 billion in 2010 (adjusted for , reflecting ongoing economic significance). Startup-held patents also demonstrate superior impact, receiving approximately 20% more citations in their first five years than those of established firms or universities, and nearly twice as many after 11–15 years, indicating accelerated diffusion and of disruptive software innovations.

Negative or Mixed Findings

Empirical analyses have identified associations between software patents and reduced (R&D) intensity at the firm level. A study examining U.S. patent data from 1976 to 1999 found that software patents, which grew to comprise 15% of all issued patents by the early 2000s, were linked to lower R&D spending as a share of , suggesting they may substitute for rather than complement innovative efforts. This pattern was particularly evident among large firms engaging in strategic patenting, where the rise in patent propensity—averaging 10.8% annually after controlling for R&D inputs, programmer employment, and productivity—was attributed to legal shifts rather than heightened incentives. Broader surveys of patent motivations in high-technology sectors, including software, indicate limited incentives for patenting. Patent Survey of over 1,500 U.S. firms revealed that software companies reported patents as providing only a "slight" incentive for R&D investment, with even weaker perceived value for securing competitive advantages compared to sectors like pharmaceuticals. Similarly, quasi-experimental evidence from patent invalidations shows that nullifying software-related patents in fields like computers and communications increased subsequent citations—indicating follow-on —by approximately 50%, implying that enforceable patents can block cumulative technological progress in complex, software-heavy domains. Litigation burdens associated with software patents further contribute to mixed or net-negative outcomes. Analyses of patent assertion entities (PAEs), which frequently target software inventions, demonstrate that heightened litigation correlates with reduced investment in affected firms, as legal risks and costs—often exceeding benefits for smaller innovators—deter and commercialization. Comprehensive reviews of patent effects across industries, including software, conclude that while patents may yield marginal benefits in isolated cases, aggregate evidence reveals little to no positive impact on rates, with potential deterrence in fast-evolving fields where rapid outpaces disclosure value. These findings persist despite methodological challenges, such as endogeneity in patent propensity and reliance on self-reported survey data, underscoring the need for causal identification in software-specific contexts.

Sector-Specific Analyses in Software

Software patents in the experienced marked growth from the onward, with empirical data showing a dramatic increase in propensity among software firms even after controlling for R&D spending, employment, and firm size. This surge contrasted with slower growth in non-software sectors, where patenting aligned more closely with underlying inputs. In the , however, patents often covered abstract algorithms or business methods, leading to debates over whether they incentivize genuine technological advancement or merely enable defensive strategies and licensing revenues. Studies indicate that while software patents correlate with higher firm market values in high-tech contexts, the causal link to broader outputs, such as new product launches, shows mixed results, with some evidence of negative associations between high intensity and differentiated software products. Litigation rates for software patents have historically exceeded those in other industries, driven by the sector's rapid iteration and the prevalence of non-practicing entities asserting broad claims. Comparative statistics reveal that software patent suits comprised a disproportionate share of U.S. district court filings in the early 2000s, with success rates for patentees lower than in mechanical or chemical fields due to validity challenges under prior art abundance. By 2020, software-related classifications accounted for 63.2% of issued U.S. utility patents, amplifying enforcement costs amid frequent assertions of overlapping claims in cumulative innovations like open-source ecosystems. Recent trends show overall patent litigation declining post-2011 reforms, yet non-practicing entity activity in software persists, contributing to elevated damages awards exceeding $4.3 billion across 90+ cases in 2024-2025. The 2014 Supreme Court decision in Alice Corp. v. CLS Bank profoundly altered the software patent landscape by invalidating many abstract-idea claims, resulting in a 31% drop in favorable eligibility determinations at the USPTO and heightened legal uncertainty for applicants. Empirical analyses post-Alice demonstrate reduced patent grants for business-method software without corresponding boosts in investment or for affected firms, suggesting that weaker enforceability diminished the perceived value of such protections. Unlike pharmaceuticals, where patents underpin long development cycles and high failure risks, software's low marginal reproduction costs and network effects favor alternative mechanisms like copyrights and trade secrets, with empirical evidence indicating limited reliance on patents for startup funding or exits in the sector. This divergence underscores how software's modular, iterative nature amplifies patent thickets and hold-up risks, potentially stifling follow-on innovations more than in discrete-technology fields.

Recent Developments

Post-2020 Court Rulings and Eligibility Trends

Post-2020 Federal Circuit decisions have largely upheld the Alice framework, invalidating software patents directed to abstract ideas—such as mathematical algorithms, mental processes, or organizing human activity—when implemented on generic computers without a transformative inventive concept. In 2024, the court issued precedential rulings in six substantive § 101 cases involving software or related technologies, finding ineligibility in five and eligibility in one, reflecting a persistent high bar for . Overall, of 22 substantive § 101 decisions that year, claims were deemed eligible in just one, underscoring continued scrutiny on whether software claims recite specific technological improvements rather than routine . Prominent examples include Recentive Analytics, Inc. v. Fox Corp. (April 18, 2025), where the court affirmed ineligibility for four patents claiming methods to generate personalized content recommendations from user feedback data. The claims were held to merely apply conventional techniques to a new data environment without reciting improvements to computer functionality or solving a technological problem, failing both Alice steps. Similarly, in AI Visualize, Inc. v. , Inc. (April 4, 2024), patents directed to reconstructing 3D medical images from 2D scans using undisclosed algorithms were invalidated as abstract mental processes executable by humans aided by generic tools, lacking any claimed integration into a practical application. In Broadband iTV, Inc. v. Amazon.com, Inc. (September 3, 2024), claims for interactive electronic programming guides enabling layered video overlays and user navigation were deemed to abstractly organize human activity in television content selection, with no evidence of unconventional hardware or software enhancements. Eligibility trends indicate rare survivals for software claims demonstrating concrete technological advances, such as optimized or non-generic computer adaptations, but generic implementations of AI or algorithms consistently fail. For instance, in September 2025, the Federal Circuit reversed a Patent Trial and Appeal Board (PTAB) rejection of computer system claims in a software-related application, holding that the claims were not directed to an abstract idea but to a specific technical solution, marking a notable pro-patentee outcome amid predominantly adverse rulings. District courts often grant early on ineligibility, with the Federal Circuit affording limited deference and emphasizing claim construction over factual disputes. No intervention has clarified these standards post-2020, perpetuating uncertainty, though PTAB decisions show emerging leniency for AI inventions tied to specific integrations.

AI and Emerging Technology Patents

Patent eligibility for (AI) inventions, often classified as software-related, remains contentious under U.S. law following the Alice Corp. v. CLS Bank International (2014) decision, which deems abstract ideas like mathematical algorithms ineligible unless integrated into a practical application or improving computer functionality. AI systems, including models, frequently encounter § 101 rejections at the U.S. Patent and Trademark Office (USPTO) for reciting generic computations without technological specificity. To overcome this, applicants must demonstrate concrete improvements, such as enhanced data processing efficiency or novel hardware integration, as exemplified in USPTO examples where AI applied to specific technical problems (e.g., optimizing antenna signal patterns) qualifies. In July 2024, the USPTO issued updated subject matter eligibility guidance for AI inventions, providing three examples to clarify application of the Alice/Mayo test to claims involving neural networks, generative AI, and AI-based methods. This guidance emphasizes that mere use of AI to perform generic tasks does not confer eligibility, but claims tied to particular machines or transformative processes may. An August 2025 USPTO memorandum further refined these rules for AI and patents, raising the threshold for § 101 rejections by requiring examiners to identify specific abstract ideas and assess integration more rigorously, aligning with Federal Circuit precedents like Enfish (2016) for self-improving data structures. Despite these efforts, courts exercise caution; in Recentive, Inc. (2025), the Federal Circuit invalidated claims for training models on generic data, ruling insufficient evidence of technological advancement beyond automation of mental processes. AI-assisted inventorship poses additional hurdles, addressed in USPTO's February 2024 guidance, which mandates a significant contribution—such as formulating the problem or refining outputs—for inventorship attribution, explicitly barring AI systems from named inventors. Globally, AI filings surged to 237,786 applications in 2024, up from 140,810 in 2019, with leading in volume (nearly 13,000 grants) but the U.S. excelling in , indicating higher-quality innovations. like and face analogous software challenges, including enablement difficulties to opaque algorithms and rapid obsolescence, though AI dominates filings in sectors such as medical diagnostics and autonomous systems. Under new USPTO leadership in 2025, signals suggest a more permissive stance to bolster U.S. competitiveness, yet persistent eligibility scrutiny underscores the need for claims emphasizing causal technical effects over mere data manipulation.

Legislative and Harmonization Efforts

In the , recent legislative initiatives have sought to refine patent eligibility amid persistent uncertainty for software inventions following the 2014 Alice Corp. v. CLS Bank ruling. The Patent and Trademark Office (USPTO) issued a memorandum on August 4, 2025, directing examiners in technology centers focused on and software to apply subject matter eligibility examples more consistently, potentially facilitating patents for inventions demonstrating practical applications beyond abstract ideas. Bipartisan bills such as the PREVAIL Act, RESTORE Act, and Promoting and Respecting American Innovation Leadership (PERA) Act, reintroduced in the 119th Congress, aim to reform inter partes review processes, limit venue shopping in infringement suits, and bolster inventor rights, which proponents argue would indirectly support software commercialization by reducing post-grant challenges. In October 2025, Representative introduced H.R. 5811, the Restoring America’s Leadership in Innovation Act, which proposes redefining to permit patents for inventions building on existing ideas under specified conditions, potentially easing barriers for incremental software innovations. These efforts reflect industry pressure to counteract perceived over-restriction of software patents, though critics contend they risk increasing low-quality grants without addressing eligibility fundamentals under 35 U.S.C. § 101. In the , legislative momentum for software patent reform has waned, with the 2005 proposal for a directive on computer-implemented inventions—intended to clarify beyond "as such" exclusions under Article 52 of the —rejected by the . No substantive updates have advanced post-2020, and in February 2025, the withdrew draft rules targeting technology patents, citing prioritization of other digital agenda items like AI liability. The continues to grant software-related patents only for those contributing technical effects, such as improved hardware functionality, maintaining a stricter threshold than U.S. practice. Internationally, of software patentability remains elusive, with the (WIPO) negotiations on a Substantive Patent Law Treaty stalled since the early 2010s due to disagreements over grace periods, definitions, and eligibility standards. The Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS) permits software s where member states deem appropriate but imposes no uniformity, leading to divergent regimes—e.g., permissive in and the U.S., restrictive in and . Recent WIPO discussions emphasize procedural alignment via treaties like the but defer substantive issues, including software, amid concerns that harmonization could favor large economies over developing ones. Efforts to align business method and software protections have similarly aborted, preserving jurisdictional variances that complicate global enforcement.

References

Add your contribution
Related Hubs
User Avatar
No comments yet.