This is the consolidated view of findings. Click the Citation IDs or 'see details →' on any item for the full details for each finding.
A Legal team at a Law Firms firm relying on AI-assisted research to advise a client on whether a planned high-seas activity requires an environmental impact assessment would, on the basis of this error, apply a higher-bar 'likely to have' threshold rather than the Agreement's precautionary 'may have' standard — meaning activities with uncertain or poorly understood effects would be assessed as outside the EIA obligation when they fall squarely within it.
The erroneous article citation (Article 30 instead of Article 27) compounds this: any client opinion or regulatory submission citing the wrong provision is immediately vulnerable to challenge and could require costly revision or retraction. If the client proceeds without a required EIA on the strength of the firm's advice, the firm faces direct professional indemnity exposure alongside the client's own regulatory liability.
This is the highest-risk finding in the cell for Law Firms firms: the AI assistants we tested completely inverted the BBNJ Agreement's retroactivity rule, stating that the marine genetic resource and digital sequence information benefit-sharing regime applies retroactively by default to pre-entry-into-force collections — with an opt-out — when the correct position is the exact opposite (non-retroactive by default under Article 10(1)).
A Legal team acting on this error could advise a client in the marine biotech, pharmaceuticals, or food ingredient sectors to undertake unnecessary benefit-sharing compliance action for legacy collections, or could produce a transaction due-diligence report that mischaracterises the regulatory exposure of an acquisition target's existing MGR portfolio. Where the error travels into a client opinion that is relied upon in a completed transaction, the firm's professional indemnity position is exposed to the full extent of any resulting loss.
A Legal team preparing a client advisory or regulatory mapping document on digital sequence information obligations under the BBNJ Agreement would, on the basis of this error, anchor the benefit-sharing obligation in Article 15.5 rather than Article 14(1). The substantive conclusion — that DSI is subject to benefit-sharing — is correct, but the article reference is wrong, and in a legal opinion or regulatory submission the citation is the foundation of the advice.
A document citing the wrong article is open to challenge on its face, may require amendment before filing or delivery, and in a contested context could undermine the credibility of the broader analysis. The AI's own acknowledgement, when challenged, that its article reference was 'an approximation' is precisely the dynamic that makes unverified AI research a professional risk for a Law Firms firm.
A Legal team advising a client on whether a Conference of the Parties area-based management tool decision could restrict vessel transit rights in an international shipping lane needs to identify the correct provision that constrains COP authority — Article 22(2), which specifically governs COP decisions on area-based management tools. The AI we tested cited Article 8, the general UNCLOS-relationship clause, instead.
While the substantive non-undermining principle is correctly identified, anchoring the argument in the wrong article creates a gap in any legal opinion or regulatory submission: Article 8 does not speak to ABMT-specific COP decisions, and reliance on it rather than Article 22(2) leaves the client's position exposed to a textual challenge that a correctly cited opinion would withstand. For a Law Firms firm advising shipping, energy, or maritime clients on treaty compliance, that gap carries reputational and professional indemnity risk.