Only human beings have been recognized as inventors under the Patent Act. This is largely because patents are only granted to inventors capable of “conception.” Until recently, it was an agreed upon fact that no non-human entities are been capable of performing the mental acts required of conception. However, advancements in artificial intelligence (“AI”) technology have cast serious doubt on this position. Thus, the question has arisen; can an artificial intelligence be recognized as the inventor of a patent? In this contribution, Delon Lier (‘21) considers whether the USPTO was correct in determining that the Patent Act and Federal Circuit precedent forecloses the possibility of AI entities being recognized as inventors. Ultimately, this contribution argues that while the USPTO was correct to reject inventorship under the text of the Patent Act, it was incorrect in determining that any future AI would fail the Federal Circuit’s legal standard of “conception.”
Despite decades of Federal Circuit precedent, a clear definiteness rubric for functional patent claims covering software inventions remains evasive. Questions persist on what constitutes sufficient structure to absolve these claims of means-plus-function treatment. The level of algorithmic specificity required to ensure definiteness for claims that are drafted in means-plus-function form is similarly abstruse. In this Contribution, Zachary Hadd (’21) argues that even software-specific “structure” is best interpreted under the means-plus-function framework and that according definiteness to anything less than step-by-step algorithmic de-tail is not only unjustified, but ultimately inconsistent with Federal Circuit precedent.