Trends and Shifts in Patenting Artificial Intelligence

Artificial intelligence (AI) is among the most essential technologies of our time, with the potential to be the next "all-purpose" technology after electricity. As AI technology progresses, AI patent activity is increasing at a rapid pace and spreading across industries.

The United States Patent and Trademark Office (USPTO) issued a new Artificial Intelligence Patent Dataset (AIPD) in July 2021. The USPTO found AI in more than 13.2 million patents and pre-grant publications in the United States, reporting more than a two-fold growth in yearly AI patent applications from 2002 to 2018.

AI in the Patent Space

AI is a broad term that covers a wide range of fields. The problem for inventors and patent attorneys is to successfully protect new AI technology development. The topic of inventorship has received a lot of attention in the debate about patenting AI innovations.

The verdict in Thaler vs. Hirschfeld by the Eastern District of Virginia in September 2021 is the first in the global controversy over AI creations, concluding that an "AI machine" cannot be an "inventor" under existing US patent rules. The Federal Circuit has since affirmed this decision.

With the rapid adoption of AI-based technologies, we may expect to see a shift in the priority of patenting AI inventions—one that guarantees that patent law's administration and handling of AI is comprehensive and flexible.

Rather than focusing on a potential expansion of patent law based on inventorship, the industry has to trend towards these three areas:

● Proactive patent examination procedures that ensure patent quality and enforceability;
● Applying existing patent laws through the lens of technology; and
● Acting with openness to new forms of IP protection to minimize disruptions to legal frameworks and promote innovation.

Examining AI Patents in a Proactive Manner

Given the fast advancement of AI technology, stakeholders must be proactive in their engagement and prudent in their consideration of methods for the patent system to support innovation. Uncertainty about the validity and enforceability of a patent might reduce its market value. To overcome this issue, we should see a trend toward prioritizing AI patent quality.

The USPTO released a report titled Public Views on Artificial Intelligence and Intellectual Property Policy in October 2020, which summarizes responses to patent-related AI questions and reflects the shift in focus to the quality, and thus enforceability, of AI patents.

In order to obtain high-quality and enforceable AI patents, written description and enablement are likely to be areas of concentration. Because of the intricacies of AI technology and the lack of transparency in how AI tools work, AI innovations provide substantial hurdles in meeting the disclosure requirement.

Many AI systems lack the capacity to describe how the technology works since the exact AI logic is unclear in some ways. Comments in the USPTO report underscore the crucial necessity for the USPTO to enforce these standards in order to ensure patent quality.

Similarly, an examination of enablement is likely, especially since enabling certain AI technologies seeking patent protection may be problematic. Such gaps in disclosure may compel the establishment of a more comprehensive AI patent disclosure system.

The amount and accessibility of prior art--or any evidence that the invention is already known-- will almost certainly be a focus. The concerns of what constitutes previous art, how much prior art there is, and how accessible prior art is will all have a substantial influence on patent quality and enforcement.

Massive volumes of prior art may be created as AI technology advances. While conventional prior art literature may describe basic AI procedures, a substantial percentage of AI technology is only documented in source code, which may or may not be public and is often regarded as difficult to find.

The USPTO is expected to fight for more resources to uncover appropriate AI-related prior art, which is required for a thorough review and the issuing of high-quality patents.

The necessity of examiner training is a recurring subject in the USPTO study. A similar drive for examiner technical training is expected, in addition to an anticipated push for the USPTO to proactively offer prior art to examiners. A more stringent obviousness criterion may be necessary at some time, but in 2022, a tactical assessment of these examination issues is the most likely way to maintain patent quality and enforceability.

There Isn't a One-Size-Fits-All Solution in Technology

Designing an AI algorithm, installing hardware to enhance an AI algorithm, or using techniques for preparing inputs to an AI algorithm, for example, all include a number of patent issues, ranging from subject matter eligibility to textual description and enablement.

The Thaler vs. Hirschfeld decision acknowledged that "as technology evolves, there may come a time when AI reaches a level of sophistication that might satisfy accepted meanings of inventorship."

Much of AI technology in 2022 will be focused on running AI models on consumer devices. This change in the AI business is due to rising privacy concerns about personal data being transported, processed, and kept on the cloud. Patenting decisions must be evaluated through the lens of the current state of technology, and improvements in AI technology should be tracked to ensure that patent interests maintain pace with AI technology developments.

Access to Other Types of IP

AI is built on the foundation of data. We may foresee a shift in attention away from merely safeguarding the AI technology, but also AI data.

AI data, including its collection and compilation, has value, and acquiring "big data" in particular may be costly. Input training data, for example, may necessitate access to millions of users' online activities and human medical data, both of which could be geographically spread and in various forms.

If present IP safeguards are unable to keep up with the rapid growth of AI technology, new types of IP, such as an IP right for nonpublic data, may be explored. This might take the form of data exclusivity rights, which are comparable to the regulatory data protection rights that apply to proprietary clinical data submitted to the FDA and other regulatory bodies.

Key Takeaways

As the field of AI continues to evolve, so will patent and intellectual property law regulating and governing it. These laws, however, must foster innovation in the field, rather than stifle it.

AI is projected to allow advances in 2022 that would be hard to achieve via human effort alone. Given the rapid advancement of AI technology, an emphasis on proactive, technology-driven, and comprehensive legal safeguards would prioritize and promote continued research in this crucial sector.