Draft AI guidelines for the Trade Marks Registry aim to curb AI-generated false citations but lack clarity on verification duties and proportionality of sanctions for unintentional errors.
Trade Marks Registry’s Draft AI Guidelines Raise Key Questions
The Trade Marks Registry (TMR) has released draft Guidelines on the Use of Artificial Intelligence in Proceedings, aiming to prevent submission of fabricated judgments and misleading evidence generated by AI tools. The move responds to increasing instances of AI ‘hallucinations’ creeping into filings, but as legal scholar Kaustub Boidi points out, the framework lacks clarity on the Registry’s own duty to verify submissions and fails to distinguish between negligent and deliberate misconduct.
The draft mandates disclosure when AI tools are used for legal research or drafting pleadings, particularly in citing case law. However, it remains unclear whether this extends to AI-assisted translation or summarization—common non-generative uses. Further, the proposed penalties, including rejection of filings or cost awards, may be disproportionate for minor or inadvertent errors, raising concerns under natural justice principles.
Practitioners should anticipate a heightened duty of due diligence when relying on AI-generated content. Until final guidelines are issued, advocates must verify all citations independently and consider maintaining internal documentation of AI use for transparency. The final policy could shape ethical obligations akin to Rule 1.1 of the Bar Council norms on professional competence.
