New Guidelines permit AI to assist patent examination, while keeping substantive judgment, verification and accountability firmly with human officers…
The Office of the Controller General of Patents, Designs and Trade Marks published the Guidelines for the Use of Artificial Intelligence in Patent Examination Procedures on August 7, 2026, laying down a framework for the use of AI in patent examination. The Guidelines recognise AI’s limitations, including the importance of prompt discipline and hallucinations, and emphasise that AI should assist, not replace, the Examiner’s or Controller’s judgment.
The Guidelines permit cautious AI assistance across a range of examination tasks, including prior art searches, preliminary assessment of sufficiency of disclosure, and improving the language of draft office communications. However, they prohibit, among other things, uploading unpublished application contents or confidential office material to public AI tools, relying on AI as a substitute for human judgment in substantive decision-making, adopting AI suggested case law or prior art without independent verification, using unreviewed AI-generated content in official communications, and relying solely on AI for decisions affecting the rights of applicants, patentees, or third parties. Also contemplated is a governance framework aimed at strengthening transparency, accountability and stakeholder confidence:
- Documentation and transparency: The competent authority may require material uses of AI to be recorded, including the tool used, purpose, date and other information necessary for supervision, audit or quality review.
- AI Governance Committee: A dedicated committee comprising representatives from the Examination, IT, and QMS (Quality Management System) Divisions is envisaged to oversee AI use, including tool approval, classification of permitted/prohibited uses, safeguards, pilot projects, quality review, policy updates, and stakeholder consultation.
- Training and capacity building: Patent Office officials should receive training on AI capabilities and limitations, confidentiality, verification of AI-generated outputs, responsible use, and identification of hallucinations or fabricated information.
- Independent oversight: Where appropriate, the Guidelines also contemplate independent audits, impact assessments, feedback systems, and incident reporting to ensure accountability and continuous improvement.
Annexure I of the Guidelines presents detailed illustrative examples of AI use in patent examination, together with the potential benefits, risks and safeguards for each use case. Annexure II provides a standard checklist and declaration for officers to record and declare the manner and extent of use of AI in their work.
Overall, the Guidelines adopt a cautious approach to AI adoption in patent examination, seeking to harness its efficiency while preserving human judgment, , independent application of mind and accountability.
The content of this article is intended to provide a general guide to the subject matter. Specialist advice should be sought about your specific circumstances.
