argument: Notizie/News - Intellectual Property Law
Source: Snell & Wilmer
On April 29, 2026 the USPTO held a webinar featuring Barry Schindler, Greg Dotson, Debbie Stevens, and Stefano Karmis that outlined how the agency is integrating generative AI across patent and trademark operations and evolving IP policy. The agency frames its AI initiatives with KPIs that include reducing pendency, improving quality, driving efficiency (including lowering expenses and reducing fraud), and increasing incremental fee collections, and has created an AI Policy Council, AI Governance Council, senior‑leader working groups, a structured approval process, and an AI risk management process. The USPTO has used traditional machine learning for years, but is now piloting generative AI: supervisory patent examiners received a USPTO‑approved generative tool in Q4 FY2025 for feedback, while patent examiners are not currently permitted to use generative AI for official examination.
Planned patent-side capabilities include drafting claim‑construction outlines, source‑anchored invention summaries to assist §§ 101 and 112 analyses, identifying priority chains for §§ 102/103, assembling claim‑to‑reference mappings, generating reference combinations and obviousness rationales, and providing § 101 eligibility scaffolds. Over the next 12 months the agency will run narrow, examiner‑designed pilots in isolated environments and will expand only if measurable quality gains appear, using metrics such as time saved, clarity of first actions, changes in after‑finals and RCEs, and examiner satisfaction; all AI output remains editable and subject to examiner judgment. Trademark operations have rapidly deployed tools: the four‑part Classification Agentic Codification Tool (Class Act) processed over 67,000 filings since its release last month and the agency launched an image search tool in early April 2026 for MyUSPTO users; examining attorneys are expected to get internal generative AI access in July 2026 after guardrails are in place. Policy developments highlighted include recent guidance on § 101 (August memo, Ex parte Desjardins), December 2025 Rule 132 guidance, November inventorship guidance, and active monitoring of disclosure requirements for AI tools/models/training data plus changes to the PHOSITA standard and related court and legislative activity. Practical recommendations in the text urge in‑house counsel to review § 101 guidance, document AI involvement in inventions, and for trademark filers to use the new tools for clearance and drafting.