The USPTO’s Guidance on AI-Assisted Inventions

The USPTO is currently consulting on its Inventorship Guidance for AI-Assisted Inventions. Here are some reflections, having participated in an expert roundtable convened by the US Special Competitive Studies Project and Center for Strategic and International Studies (CSIS). 

The discussion was under Chatham House rules, so here are my thoughts:

• Guidance on AI-assisted inventions is welcome but we need legislation and, ideally, internationally harmonised approaches. Patent protection is socio-economic engineering and should be shaped by lawmakers in consultation with stakeholders and underpinned by economic analysis.
• The USPTO suggests that the patent system is designed to encourage human ingenuity. But is that an end in itself or merely the historical de facto route to the ultimate goal of “progress of sciences and the useful arts”? If “progress of sciences and the useful arts” is the ultimate aim, obviousness seems a better filter than excluding patents based on the level of human contribution.
•  The USPTO suggests that the principles of joint inventors should be applied to AI-assisted inventions to check for “significant contribution” by a human. To a non-US practitioner, the USPTO’s summary of the law seems to raise but not explore two aspects of US legislation/caselaw. First, the definition of inventor applies, inter alia, to someone who “discovered” the subject matter of the invention. Second, the USPTO rules out “unrecognized accidental creation” but does not (on my reading) explain why recognized accidental creation is insufficient. The USPTO’s third principle seems to elide reduction to practice with recognition: “Reducing an invention to practice alone is not a significant contribution that rises to the level of inventorship. Therefore, a natural person who merely recognizes and appreciates the output of an AI system as an invention, particularly when the properties and utility of the output are apparent to those of ordinary skill, is not necessarily an inventor.” (Emphasis added)
• Is caselaw on “joint inventors” the best analogy? Where humans are vying for sole/joint rights to an invention (or a copyright work), the Court sets a high threshold for individual contributions. But is that the case where rights are not contested (it does not appear to be the case in many copyright cases).
• Should we prevent “push-button” invention because, e.g., there is a low marginal cost for each invention? First, patent law does not consider the cost of invention (e.g. whether an invention “resulted from long toil and experimentation or from a flash of genius”) so why not invention by prompt? Second, the costs of patent prosecution and maintenance will prevent a flood of push-button patents – even the largest companies must be selective in the patents they pursue.
• The guidance may invite analysis of whether prompts reach the threshold of significant contribution. This may be complex and arbitrary.
• Whatever the approach to patents per se, we must ensure incentives to commercialise inventions that are costly to bring to market, such as new pharmaceuticals.

The Guidance does not have the effect of law but does apply to the examination of all patent applications. It will be interesting to see if examiners begin to explore the use of AI during examination.