Artificial Intelligence (AI) Patents

Data from granted patents, patent applications and scientific publications indicate growth in artificial intelligence (AI) patents.

Patent Eligibility

One of the bigger challenges facing the patenting of AI and machine-learning inventions is subject matter eligibility under United States Code 35 § 101, specifically applied in Alice and its descendants. While the case law since Alice has provided incremental guidance, fundamental uncertainties with this test still remain. There is no clear test on what constitutes an “abstract idea” and no clear test on when additional claim elements constitute “significantly more” than an abstract idea.

Nevertheless, the Federal Circuit has provided some guidance in cases such as Enfish, McRO, DDR Holdings, Bascom, Amdocs, Trading Technologies International, and others, for navigating the abstract-idea rabbit hole of Alice for computer-implemented inventions. For example, the Federal Circuit has clarified that in the step-1 inquiry, it must be determined whether the claim as a whole is directed to an abstract idea.

This guidance is helpful because the USPTO had been, in instances, applying step 1 to inquire merely whether a claim recited an abstract idea (which, at some level of paraphrasing and generalization, can be found in almost any claim). This clarification helps patent applicants argue that a claim as a whole is not directed to an abstract idea (step 1) at all, potentially avoiding step 2 scrutiny altogether.

The Court has also indicated that claimed inventions that represent technological solutions to technological problems, and claimed inventions that are rooted computer technology, may have a good basis for being patent eligible.

Patent Disclosure

Challenges existing in the area of disclosure adequacy are wide and deep, covering at least two major topics—claim scope and transparency. While disclosure would generally cover what was known at the time of training, a broader scope of claim may be needed to encompass what the AI enabled invention does.

In fact, understanding exactly how the AI enabled invention gets to the end-result is problematic (e.g., this is analogous to a teacher admonishing a student for having the right answer without detailing how they got it). Additional problems may arise from the use of privacy restricted data pools in model training or deployment. For example, under current practice, written description support for computer-implemented inventions generally require sufficient disclosure of an algorithm to perform a claimed function, such that a person of ordinary skill in the art can logically decide that the inventor had ownership of the claimed invention.

Enablement Requirement

Additionally, the USPTO has posited several questions regarding how existing statutory requirements apply to artificial intelligence inventions. This category of questions presents fewer problematic issues than inventorship and ownership. The USPTO’s interest includes comments directed towards how existing rules might apply in AI-contexts, such as:

• The written description (to show that the inventors had possession of the invention at the time of filing) and enablement requirement (to teach how to make and use the invention);
• Obviousness; and
• Subject matter eligibility.

With respect to written description and enablement, AI-based inventions may make use of models or networks that evolve as part of a training process that is probabilistic in nature, making reproducibility challenging or impossible.

AI Impact

To evaluate the adequacy of a patent’s specification for an AI technology therefore requires consideration of the technologies being claimed, and must be adequate to enable one of ordinary skill in AI technology to prepare the claimed AI invention.

Prior Art Requirement

Current law generally limits the scope of the relevant art to analogous prior art and inventors can rely on this to argue against combinations that are alleged to be obvious. However, with an Inventive AI, “the scope and content of art” may be expanded to go beyond analogous art, given that the Inventive AI would most likely be presumed to be capable of knowing and processing virtually all of the prior art, at least all that is online or otherwise accessible.

Patent Claims

Claiming a deep machine learning algorithm’s application to any type of data may require disclosing enough species of data set types to which the algorithm is applied to thereby enable the genus of all dataset types.

Thus, to satisfy § 112(a), the data set used for training the learning algorithm (be it an inference machine, binary logic tree, or neural network) will need to be disclosed in the application to enable others to use the same or similar data sets to practice the claims. On the one hand, given the central nature of the training dataset, if only one data set is disclosed, the claims will necessarily be limited to that dataset.

If the applicant seeks to broaden the claim scope, more than one data set will need to be disclosed. However, a data set that is merely more of the same type of data will not suffice in supporting broader claims under § 112(a). On the other hand, a clearly different data set will train the learning algorithm to do a different thing.

Thus, unless there is a way to show that two different data sets can be used to train the AI system to do the same thing (i.e., the claims of the invention), the disclosure requirements under 35 U.S.C. § 112(a) are likely to limit the scope of the claims to the specific disclosed training data sets and their respective disclosed training outcomes. In order to obtain claim scope that encompasses additional training outcomes, the specification will likely need to further disclose the training data sets used to obtain the desired training outcomes.

Accordingly, should the goal of an AI patent application be to support the broadest AI invention claims possible, the adequacy of the disclosure will likely require disclosure of many examples of application of the AI invention.

Other issues regarding patenting AI Inventions At the present time, in the U.S., there are no specific judicial decisions or statutory laws to address these potential problem areas in the non-obviousness standard for Inventive AIs, but these will become issues if Inventive AIs become recognized as inventors in the future.

AI Patent Challenges

Humans remain essential in monitoring and developing AI, and technology is relevant as long as it promotes transparency and objectivity for the public. Thus, the overall purpose of AI is to facilitate the process of litigation for advocates and judges rather than displace them. With this, there come substantial challenges. The biggest challenge is for users to get accustomed to the AI as it has not made any remarkable inlay into the patent laws.

Considering other countries, they embarked on the journey to e-discoveries much before India was introduced to the concept of AI itself. e-Discovery is the electronic component of recognizing, gathering, and presenting electronically stored information or data (ESI) at the behest of a request for presenting the same in the matter of a lawsuit, case, investigation, or examination.

Thus, it can be used as evidence in the Civil or Criminal cases, and also be employed to draft watertight contracts, reviewing legal documents, etc. Software like this is used as a second set of eyes to recheck important content and identify if anything is missing or incorrect. However, this is yet to be adopted to India which is a considerable challenge because a significant segment of the Indian population still lacks the basic understanding of the digital spectrum, be it computer usage, online payments, online filing, e-courts, etc.

Taking these aspects into account, AI is a breakthrough when we consider Intellectual Property (IP) and Software Patent Rights. AI Software has made a huge difference in the efficiency of this sector of law with patent filing, searching for existing claims, granted trademarks, copyrights filings, infringements, etc, as it makes these processes much easier and less time-consuming. This has also been witnessed in Contract Law. Thus, AI is building a lot of opportunities, but it will take time, and for it to spread effectively users must incorporate these updates into our day to day lives.

Patent attorneys work closely with clients to ensure creation of successful AI patent applications to protect the business goals of the clients. We navigate the domestic patent and international patent processes, assisting clients with efficiency, by drafting and filing utility patents, design patents, conducting patent searches, and, preparing office action responses.

We handle effective patent prosecution to ensure that patent is granted swiftly, and after the patent application has been approved, we work with clients to license, transfer and sell their patent. We are adept at helping clients implement a strong plan to monetize and protect their inventions through Patent Licensing, Patent Manufacturing agreements, and Patent Nondisclosure agreements. We also assist clients with trademark services, including trademark searches and trademark filing.

Law Office of Patent Attorney Rahul Dev offers high value software patent drafting and patent due diligence services to clients by using proprietary and efficiently proven process along with a fixed fee costs, for performing comprehensive patent investigations and providing clients with strong patent reports for decision making.

We provide comprehensive Patent and Trademark legal services via our global network to create valuable patent portfolios and resolve complex patent disputes by providing patent litigation support services.

Our team of advanced patent attorneys assists clients with patent searches, drafting patent applications, and patent (intellectual property) agreements, including licensing and non-disclosure agreements.

Advocate Rahul Dev is a Patent Attorney & International Business Lawyer practicing Technology, Intellectual Property & Corporate Laws. He is reachable at rd (at) patentbusinesslawyer (dot) com & @rdpatentlawyer on Twitter.

Quoted in and contributed to 50+ national & international publications (Bloomberg, FirstPost, SwissInfo, Outlook Money, Yahoo News, Times of India, Economic Times, Business Standard, Quartz, Global Legal Post, International Bar Association, LawAsia, BioSpectrum Asia, Digital News Asia, e27, Leaders Speak, Entrepreneur India, VCCircle, AutoTech).

Regularly invited to speak at international & national platforms (conferences, TV channels, seminars, corporate trainings, government workshops) on technology, patents, business strategy, legal developments, leadership & management.

Working closely with patent attorneys along with international law firms with significant experience with lawyers in Asia Pacific providing services to clients in US and Europe. Flagship services include international patent and trademark filings, patent services in India and global patent consulting services.

Global Blockchain Lawyers (www.GlobalBlockchainLawyers.com) is a digital platform to discuss legal issues, latest technology and legal developments, and applicable laws in the dynamic field of Digital Currency, Blockchain Patents, Bitcoin, Cryptocurrency and raising capital through the sale of tokens or coins (ICO or Initial Coin Offerings).

Blockchain ecosystem in India is evolving at a rapid pace and a proactive legal approach is required by blockchain lawyers in India to understand the complex nature of applicable laws and regulations.