Draft guidelines from the Trade Marks Registry to regulate AI-generated evidence fail to define verification obligations and proportionate penalties for submissions involving fabricated judgments or AI-designed marks.
Trade Marks Registry’s AI Guidelines Fall Short
The Trade Marks Registry issued draft guidelines in October 2026 to regulate the use of artificial intelligence in proceedings, aiming to curb fabricated citations, AI-generated designs, and misleading translations. While the move responds to genuine concerns — including instances of AI hallucinated case laws and synthetic mark submissions — the framework lacks clarity on verification duties and proportionality in sanctions.
The guidelines require declaration of AI use but do not mandate verification mechanisms for authenticity of cited judgments or source data. Nor do they distinguish between permissible AI-assisted drafting and wholly AI-generated evidence. This creates ambiguity for practitioners using AI tools for translation, proofreading, or prior art searches.
"The burden of truthfulness remains on the advocate,"the guidelines state, but stop short of prescribing audit trails or technical validation.
For trademark litigators, the absence of risk-tiered consequences — e.g., distinguishing minor formatting errors from deliberate falsification — may lead to inconsistent enforcement. The final guidelines must clarify thresholds for misconduct and align with emerging judicial standards on digital evidence integrity.
Citations
- Trade Marks Act, 1999
- Bar Council Rules, Professional Conduct
