Lawyers
Litigation associates Rana Matared and Harut Minasian published an article in the Chicago-Kent Journal of Intellectual Property examining the nuances of patent eligibility for “ethical AI” inventions. In the article, “The Patentability of Ethical AI: A Spectrum,” published on February 18, the authors write that the patent eligibility of inventions for promoting ethical AI fall on a spectrum. “At one end of the analytical spectrum are result-oriented claims—generally indicative of patent ineligibility—that merely state goals (e.g., “be fair,” “reduce bias”) or automate human review using generic computing resources,” the authors say. “At the other end of the spectrum are concretely specified systems that alter how computers process, store, and act on information,” including but not limited to “fairness checks embedded into the AI model training process to reduce biased outcomes,” they note.
Rana and Harut add that the law will slowly catch up with technology. “Consistent with the centuries-old progress of the common law, patent eligibility of AI technology will evolve slowly over time as the technology matures and the law of patent eligibility develops accordingly.”
» read the article
Related Insights
