“Patentability of AI”

Submitted by:-

Ananya Tripathi


University of Petroleum and Energy Studies, Dehradun.


The junction of conventional patent concepts and the distinctive features of artificial intelligence gives rise to the complicated and dynamic field of patentability in intellectual property law. Below is a synopsis of the main ideas-:

  1. Define and Extend:- Artificial intellect is the term for computer systems which created to carry out activities like speech recognition, visual perception, language translation, and decision-making that normally need human intellect, although determining if an AI innovation is patentable entails assessing whether it satisfies the requirements for patent protection, which includes uniqueness, non-obviousness, and utility.
  1. Requirements for Patent Qualification:- The AI creation needs to be fresh and unreported in earlier works of art. On the other hand, an innovative step must be involved in the invention for it to be apparent to a person competent in the field. Moreover, the invention must be practical and have a clear, significant, and reliable usage for it.
  1. The difficulties in obtaining an AI patent:- It’s difficult to identify the creator of an innovation produced by AI. Inventions are traditionally patented by humans, although artificial intelligence systems are capable of producing novel solutions on their own. Machine learning-based AI systems in particular can be “black boxes” with mysterious inner workings whereas, it makes difficult to meet the requirements of providing a sufficient level of disclosure for the innovation. And a lot of inventions pertaining to AI can be regarded as abstract notions, which are not currently protected by the patents. Mental processes, mathematical techniques, and algorithms fall under this category.
  1. Jurisdictional Variations:- Different jurisdictions have different methods and criteria regarding AI patentability whereas, this raises legal and policy considerations. There might be differences in the interpretations and guidelines between the ‘United States Patent and Trademark Office’, the ‘European Patent Office’, and other international entities. Artificial Intelligence as Inventor. The question of whether AI systems may be included as inventors on patent applications is still up for dispute. There are plans to modify several legal systems to allow for inventions produced by AI. The patentability of artificial intelligence gives rise to a number of ethical and social implications, including how patentability may affect innovation, how best to strike a balance between defending the rights of inventors and facilitating public access, and whether or not monopolies on core AI technologies are possible.
  1. Current Progress:- AI patentability has been the subject of court cases in a number of nations, which have established precedents that may influence future patent applications and laws. To address the particular difficulties presented by AI in the patent system, many jurisdictions are contemplating or have already enacted legislation. To guarantee a logical and uniform strategy, efforts are being made to standardise AI-related patent laws among various nations.

Criteria for Patentability:-

AI-assisted inventions, like any innovation, must still fulfil certain requirements in order to be eligible for patent protection, even if artificial intelligence is not yet capable of being an inventor. Here are the major ones that are as follows-:

1. Novelty: The invention must be fresh and unknown to the general public. This implies that before the patent application is submitted, it cannot have been mentioned in a previous patent, publication, or public usage.

2. Non-Obviousness: An individual with sufficient knowledge in the relevant subject cannot find the innovation immediately apparent. It should be a substantial improvement above what is already available and not something that could have been readily worked out by a layperson in that subject.

3. Industrial Applicability: The innovation needs to be able to be produced or applied in an industry, which means that it cannot be an abstract concept or merely theoretical based.

4. Statutory Subject Matter: The innovation must fit into the lists of items that qualify for patents in some nations.  Generally speaking, this leaves out things like natural facts, rules of nature, and abstract concepts.

The Uncharted Territory of AI Patentability:-

Artificial intelligence poses an intriguing challenges to intellectual property law which especially with regard to the notion of patentability. Historically, those who have invented new and original solutions to technological issues have been granted patents. As AI develops, though, the question of whether AI can create innovations on its own or if those created by AI may be patentable emerges.

Intellectual Property Tools to Safeguard AI:-

Inventors could legally bar others from producing, using, importing, or selling their inventions for a set amount of time by obtaining a patent. The “Indian Patents Act, 1970”, specifically prohibits the patentability of “mathematical methods, business methods, computer programs per se, and algorithms” whereas, this ‘section 3(k)’ of the act defines the “subject matter exclusions that are applied to AI-related innovations in India”, India grants patents to AI-based ideas that are new, practical, and non-obvious. On the other hand, trade secrets are exclusive details that provide the owner with a competitive financial edge. Information include client lists, company strategies, financial data, and technological data are examples of what it may contain. Trade secrets are safeguarded in India under the “Indian Contract Act of 1872”. Trade secret protection can be used to stop others from utilising or exposing private data, which is frequently included in AI-based innovations in the form of data sets.

Inventive AI:-

The progress made in AI has been transformed machines from simply tools for creativity to significant contributors to creation. The medical field is using AI machines to find new drugs. Microsoft is developing a system known as “Hanover” that would house all the information on medications used to treat cancer. By utilising this information, the system will be able to forecast which combination of medications would be most effective for a given patient’s diagnosis. These devices, dubbed innovative artificial intelligence, generate original works with little to no assistance from humans, though, people investigate two well-known creative AI computers that have been assisting in the creation of new technologies across the country. Created by Dr. Stephen Thaler, the “Creativity Machine” has been producing its own works of art since 1994. The device “came to the closest yet to emulating the fundamental neurobiological mechanisms responsible for the formation of ideas”. A set of on/off switches that automatically link to construct software without the need for human interaction is called an artificial neural network, and it is a feature of the machine. First, the artificial neural network feeds the computer with a vast quantity of data, and it automatically determines which of the data is suitable for developing new content and which are not. The machine began to generate 11,000 new songs in a single weekend after Dr. Thaler programmed it with his favourite music. The machine produced not just music but also a cross-bristle pattern.

Challenges of Patentability of AI:-

The traditional patent system is put to the test by the rapidly developing areas of artificial intelligence, which raises a number of issues with patentability-:

  • The Inventory Dilemma: The focus of modern patent law is on human inventors. As evidenced by the ‘Thaler v. Vidal’ case in 2022, it is difficult to identify an AI system as an innovator.
  • Indefinable Originality: The question of whether AI is genuinely innovative or merely recombines preexisting knowledge arises because AI frequently learns from large databases. It could be more challenging to show innovative steps ie.,non-obviousness.
  • Black Box Issues: Certain AI systems, especially those that are sophisticated, have illogical thinking. This hinders the patent disclosure requirements by making it difficult to describe how the AI came up with the innovation.
  • Indistincting between Patentable Subject Matter: With AI, it might be difficult to distinguish between technically patented solutions and unprotect-able abstract concepts or algorithms. It might be challenging to separate the technological role that AI plays in inventions.
  • International Discrepancies: Every nation has a different patent law. In Europe or China, there may be more stringent restrictions on software or algorithm patenting than there are in the US.
  • Potential for Abuse: Artificial intelligence patents that are excessively broad and may hinder progress by limiting access to necessary resources. Achieving the ideal balance between innovation protection and development promotion is essential.
  • Unforeseen Moral Issues: Ethical questions about ownership and management of innovations produced by AI may surface as its capabilities grow. Who has responsibility for the unfavourable effects of AI innovations.

Case laws:-

  1. DABUS Case-:

A well-known example is the artificial intelligence system called as ‘DABUS’-Device for the Autonomous Bootstrapping of Unified Sentience. Whereas; Dr. Stephen Thaler, the man behind DABUS, has registered a DABUS as the inventor on patent filings in a number of nations. Moreover; the feedback has been inconsistency:-

United States- Only natural beings may be designated as inventors, according to decisions made by the USPTO and a US district court.

Europe- The European Patent Office (EPO), which upholds the need for a human inventor, likewise denied the application.

Australia- The Federal Court first declared that DABUS was an inventor; however, the Full Federal Court later reversed this ruling.

South Africa- Approved a patent with DABUS listed as the inventor, which is a notable departure from other countries.

  1. Punjab State v. Jaswinder Singh:-

In this case, it is held that a person received an actual term of 10 months and 21 days, out of the entire 2 year sentence that was issued. Although, the Court believes that it would be just if the petitioner’s two-year term of hard imprisonment which were reduced to the time a person has already served.

  1. The 2017 case of Justice K.S. Puttaswamy (Retd.) v. Union of India:-

This case deals with the nine-judges bench unanimously upheld the right to privacy as a basic freedom guaranteed by the Indian Constitution. The Court decided that the right to privacy was a basic component of liberty, autonomy, and dignity and that it was essential to the freedoms protected by all fundamental rights.


To the best of my knowledge, the field of artificial intelligence patentability is developing, and it is necessary to strike a balance between the requirement for transparent, equitable, and flexible patent rules and the safeguarding of cutting-edge innovations. The future terrain of AI patentability would be shaped by the ongoing legal, technological, and ethical disputes. Moreover; Patent law differs from nation to nation and could be complicated, although this is a broader view, which is strongly advised that people could speak with a patent counsel to find out whether an AI innovation is patentable or not. Therefore; late changes in AI’s patentability are indicative of a continuous endeavour to modify conventional patent systems to suit the distinct difficulties presented by AI innovations. A more complex and all-encompassing approach to AI patentability is being influenced by developments in technology, policy debates, legislative changes, and court rulings. To guarantee that patent systems efficiently stimulate innovation while ensuring justice and transparency, ongoing conversation and modification are important while innovation is upholding justice and more clarity to it. At last, The above challenges or difficulties call for continued debate and possible changes to the patent law system.


  1. Is It Possible for Machines to Invent?
  2. Are Machines Capable of Being Inventors?
  3. What do you understand by “Artificial Intelligence”?




Leave a Reply

Your email address will not be published. Required fields are marked *