Source: WebProNews

Source: WebProNews. Silicon Valley and the pharmaceutical industry are adapting to a new "patent playbook" following the USPTO's strict guidance on AI inventorship. While AI cannot be an inventor, the new framework allows humans to patent AI-assisted outputs if they can prove they made a "significant contribution" to the conception. This has led to a new legal defense strategy where researchers must meticulously document their "domination" over the AI tool. Companies are now maintaining detailed "invention logs" to distinguish between AI suggestions and human decisions, ensuring they don't treat the AI as a "black box."

The article explains that the USPTO is applying the "Pannu factors"—a legal test from 1998—to AI scenarios. To qualify for a patent, a human must do more than just prompt a system; they must understand the underlying mechanism and contribute to the specific structure of the solution. This creates a complex bureaucratic burden but is seen as essential to prevent R&D capital from fleeing to jurisdictions with looser rules or disappearing into trade secrecy. The guidance aims to thread the needle between discouraging "junk patents" generated by bots and encouraging legitimate, AI-accelerated scientific inquiry.