Is AI Making Patent Prosecution Better? (with Diego Freire)

AI can draft a patent application in minutes. But speed isn’t the same as quality, and in patent prosecution, the difference matters enormously.

Diego Freire breaks down the risks practitioners and IP teams need to know about AI in patent prosecution, including:

• Hallucinations pose a real risk, as fabricated prior art citations and imprecise claims language can implicate the duty of candor to the USPTO
• “AI slop” (generic, low-precision output) is a particular hazard in a field where exact language determines scope and enforceability
• Training data limitations mean AI tools may not reflect current examination practice or recent Federal Circuit decisions

Used thoughtfully, AI can add significant value by saving you time on prior art searches, spec drafts, and routine tasks. Those getting the most out of AI are those who treat it as a tool, not a replacement for the judgment and client understanding that experienced attorneys bring.