Comments to the USPTO regarding AI and inventorship

More than 50% of the patents granted in 2020 were related to AI, according to a USPTO report. Considering this large volume of AI-related patents, the office recently sought feedback regarding AI inventorship. Key points made by AUTM, BIO, Google, IBM, and Microsoft are summarized here.

AUTM made the following statements in response to USPTO’s comments:

• AI inventions must be protected in a way that encourages both innovation and disclosure.
• Current law can adequately address AI innovations, and it is unnecessary and perhaps counterproductive to change the law based on speculation about the future of AI.
• Rather than recommending changes to the patent laws, AUTM supports guidance that reinforces the patentability of AI-enabled inventions and inventions on novel AI architectures.
• AUTM also supports improved protection of the necessary datasets to further AI innovation.

The Biotechnology Innovation Organization (BIO) noted that the use of artificial intelligence and machine learning (together: “AI”) tools is becoming increasingly common among their member companies. The organization’s comments included:

• BIO members view AI as a “tool” that facilitates human invention-making.
• AI is not deemed to possess the agency, purpose, motivation, or the capacity for ideation that is required to establish conception of an invention.
• The use of an AI tool, even one that significantly facilitated the conception of an invention, should not relegate a human inventor to the status of part-inventor, co-inventor, or non-inventor.

In its introduction to their comments, Google encouraged USPTO to “shed light on inventorship,” noting that “inventorship can be a challenging area for patent applicants as it is highly fact dependent and often complicated.” IBM also encourages the USPTO to implement mandatory technical training for any patent examiner who is examining AI-related innovations, stating, “As the USPTO’s Inventing AI report demonstrates, the number of patent examiners who are now examining these AI-related inventions makes up a substantial portion of the examining corps, and that number is only going to increase. … We urge that a robust technical training program be put in place so that patent examiners are well-situated to assess whether or not to grant a patent.”

IBM made the following selected statements in response to USPTO’s questions:

• Requiring applicants to explain all contributions by AI systems would place unnecessary burdens on applicants and the examination process.
• There is no single definition of AI. If one assumes that AI is anything automated that mimics a human capability, it could cause applicants to unduly capture information beyond the scope of what the patent statute requires.

Microsoft emphasized that AI tools are, in all cases, human-directed tools. They also noted that the AI tools are fallible, so humans must “validate and revise” to ensure a useful end product. They also stated, in response to specific USPTO questions, that:

• Using an AI system is not fundamentally different from using other technical tools in the invention-creation process.
• The people who conceive of an idea and recognize and verify its value, whether aided by tools or other humans, are inventors. … Microsoft does not believe that AI contributions rise to the level of inventorship and, therefore, do not raise ownership issues.

Go here for all of the responses submitted.