The China National Intellectual Property Administration (CNIPA) has announced significant revisions to the Patent Examination Guidelines, effective January 1, 2026.

This alert summarizes the critical regulatory changes—including AI inventorship bans and stricter dual filing rules—and outlines strategic drafting methodologies to navigate this evolving landscape.

1. AI Inventors Explicitly Prohibited
The revised Guidelines (Rule 14) mandate that an inventor must be an individual. Entries such as “AI” or “Project Group” are strictly forbidden. Patent agencies are now required to verify the authenticity of applicant identity in-formation.

2. Mandatory Claiming of Inventive Features
New updates to Part II, Chapter 4 strictly require that features contributing to inventive step (e.g., unexpected technical effects) must be explicitly recited in the claims. Examiners will no longer consider features solely disclosed in the specification when evaluating inventiveness.

3. Clarified Eligibility: AI & Data Compliance
Patentable: AI solving technical problems is eligible (e.g., neural networks for grading scrap steel).
Unpatentable: Inventions violating social ethics or the Personal Information Protection Law are rejected. Examples include algorithms for covert data collection or unethical autonomous decision-making.

4. Strategic Pivot: “Response Ability” in Drafting
To counter stricter examination, drafting must shift from “passive recording” to “active construction.” This involves forward-looking layouts that predict examination disputes (novelty/inventiveness) and embed “response anchors”—specific fallback positions in dependent claims.

5. New “Either-Or” Rule for Dual Filings
A major shift for “double insurance” strategies: if a declared invention application overlaps with a granted utility model, the applicant must abandon the utility model to secure the invention patent. The previous option to simply amend the invention claims to coexist has been removed.