<- Take me back to my Legal x Corporate Banking (SG) overview
Executive Summary
In-house legal teams at Singapore corporate banks advise on capital instrument issuance, group-structure transactions, capital adequacy compliance disputes, and the legal characterisation of MAS amendments to Notice 637. Across the two findings in this cell, an AI model fabricated a parallel holding-company notice and misrepresented MAS's amendment-PDF editorial convention. For corporate banking legal teams, both failures generate direct legal advice risk: legal positions documented against a fabricated notice or an editorial-annotation reading the regulator's own text contradicts.
How AI gets this regulation wrong
Both findings are inference drift. The AI produced confident specific answers on instrument identification and on amendment-convention reading, in cases where the regulator's published text resolves the question. The model committed where the correct posture would have been retrieval of the actual notice paragraph or a 'cannot verify' response.
| AI's Failure Mode | Count | Affected findings |
|---|---|---|
| Exposed Fabrication | 2 | Finding#1 · Finding#2 |
What that means for your practice
For corporate banking in-house counsel, both findings translate into legal advice risk where the legal position is captured against the wrong regulatory basis. The risk is concentrated in capital instrument structuring advice, group-perimeter legal opinions, and the legal interpretation of MAS amendments where the bank's legal function is the first reader of the published text.
| Risk Impact | Count | Affected findings |
|---|---|---|
| Regulatory enforcement / professional liability exposure | 2 | Finding#1 · Finding#2 |
When this affects Legal teams at Corporate Banking firms
Corporate banking legal teams apply Notice 637 across capital instrument structuring (issuance documentation for Tier 1 and Tier 2 instruments, prospectus disclosure on regulatory capital treatment), legal opinions on group-level capital adequacy structuring (particularly where the corporate bank sits within a Singapore financial holding company), advice on supervisory engagement with MAS, and legal characterisation of amendments to Notice 637 for the bank's compliance and reporting functions.
The two findings in this cell map onto two question types that recur in this work. First, instrument identification: when advising on group structure or capital adequacy at the holding company level, what MAS instrument is the source of obligation. The AI's fabrication of a parallel notice produces an opinion-grade citation that does not exist. Second, amendment-convention reading: when interpreting MAS's tracked-change amendment for the bank's compliance and reporting functions, how to read the yellow highlighting.
The AI's wrong characterisation causes the legal team to treat editorial annotations as live regulatory text, embedding the error into legal advice on the amendment's effect.
The findings at a glance
The table below lists each finding from the AI testing on MAS Notice 637 in this cell, showing the topic, the AI's failure mode, and the citation identifier.
Aggregate impact
Taken together, the two findings describe an AI generation pattern that corporate banking legal functions should expect on Singapore prudential work. The model is willing to commit to a fabricated instrument and to an incorrect reading of a regulator-specific convention; both outputs are confident and silent.
The downstream legal consequences are immediate. A capital instrument issuance opinion that recites a fabricated parallel notice misstates the regulatory characterisation; supporting prospectus disclosure carries the error into the public document. A group-structure legal opinion built on the fabricated instrument misstates the perimeter on a question the source notice resolves through its own scope rule. A legal characterisation of the 2025 amendment built on the AI's yellow-highlight reading treats annotation text as live regulatory obligation; legal advice flowing from that characterisation misdirects compliance and reporting work on the amendment.
What your team should do
In-house legal teams at corporate banks should treat AI tools as research-prompt generators on Notice 637 work, with a mandatory verification step against MAS's published text before AI output enters a legal opinion, prospectus disclosure, or supervisory submission.
Practical safeguards: (a) every MAS instrument citation in a legal opinion or transaction document must be matched to the MAS publications portal listing. (b) When characterising a MAS amendment's tracked-change conventions for the bank's downstream functions, pull paragraph 3 (or the equivalent reading-convention paragraph) of the amendment before issuing advice. (c) For group-perimeter capital adequacy advice, anchor the legal position on Notice 637 paragraph 11.2.2 (the scope rule), not on AI-supplied parallel-instrument references. (d) Build a legal-function carve-out into the bank's AI-use policy for instrument-identification and amendment-interpretation queries, where this testing shows the AI is confidently wrong.
Where AI tools deliver value in legal work on this notice: outlining capital instrument opinion structures, identifying which Parts of Notice 637 are likely relevant to a particular structuring question, drafting client-facing summaries of regulatory architecture for review against source, and surfacing cross-references between Notice 637 and adjacent MAS instruments for verification.
How RLB Can Help
RegLeg's published Hallucination Research is available as a free pre-flight check for Singapore banking practitioners working across MAS-supervised entities. Before relying on AI-assisted output for regulatory interpretation, compliance advice, or capital-instrument structuring, practitioners can consult the research to identify where AI tools have demonstrably mis-stated the rules: invented instruments, misread editorial conventions, outdated paragraphs presented as current. The research covers specific MAS instruments and surfaces the exact questions where AI tools have failed, making it a practical reference rather than a general caution.
For firms where multiple teams are working the same regulatory portfolio, RegLeg offers bespoke deep-dives into individual MAS instruments. These engagements go beyond the published findings to examine the full pattern of AI failure modes relevant to the instrument: the question types, the failure mechanisms, and the risk implications for compliance, risk, treasury, legal, and reporting work. The output is designed to be shared across functions and used as a durable reference, reducing duplicated due-diligence effort and creating a consistent internal standard for AI-assisted regulatory work.
RegLeg also develops training and CPD-aligned content for Singapore banking teams. The material translates the failure-mode catalogue into practical guidance on the classes of error practitioners should watch for: confabulated cross-references, version confusion between superseded and current instruments, jurisdiction bleed between superficially similar regimes, and inference-driven elaboration that overstates what an instrument actually requires. Separately, RegLeg offers a confidential review of a firm's existing AI-use policy against the failure-mode catalogue, identifying gaps between the policy's assumptions and the documented evidence of how AI tools perform on Singapore prudential questions in practice.