AI Hallucinations Affecting Governance & Company Secretarial Teams at Payment Institutions Firms in international jurisdictions
This page aggregates AI hallucination findings affecting Governance & Company Secretarial teams at Payment Institutions firms in international jurisdictions across 1 regulation(s).
Findings overview
| Regulation | Hallucinations | Blind spots | Total |
|---|---|---|---|
| Principles for Financial Market Infrastructures (PFMI) | 1 | 1 | 2 |
| Total | 1 | 1 | 2 |
Principles for Financial Market Infrastructures (PFMI)
Hallucinations (1)
Fabricated PFMI board risk committee citation
-
AI's failure: AI confidently answered incorrectly; when challenged, it admitted it didn't really know — right or wrong
-
Risk for Governance & Company Secretarial at Payment Institutions: Companies registry or listing authority action against the Payment Institutions entity
When a Governance or Company Secretarial team at a Payment Institutions firm asks AI tools whether PFMI Principle 2 mandates a risk committee, the AI assistant we tested cited a specific key consideration sub-number and produced quoted regulatory language that does not appear in the source document — presenting a conditional recommendation as though it were a settled textual standard.
If the team incorporates this fabricated citation into board terms of reference, a governance framework review, or a response to a regulatory enquiry, the firm is relying on invented authority for a board-level governance decision. CPMI and national overseer examinations of FMI governance arrangements routinely scrutinise whether committee structures are supported by the applicable international standard; a framework built on a non-existent sub-provision will not withstand that scrutiny.
The cost of unwinding a governance error at board level — re-opening approved terms of reference, correcting regulatory submissions, and explaining the discrepancy to overseers — substantially exceeds the cost of source verification at the drafting stage.
Blind spots (1)
Blind spot on IOSCO disclosure framework verbatim content
-
AI's failure: AI couldn't find the real answer even with web search enabled
-
Risk for Governance & Company Secretarial at Payment Institutions: Board pack, statutory filing, or AGM resolution rests on a procedural rule that doesn't exist
When a Governance or Company Secretarial team at a Payment Institutions firm asks AI tools for the precise verbatim text, thresholds, or cross-references from the IOSCO co-published PFMI disclosure framework and assessment methodology, AI assistants we tested correctly declined to fabricate content — but were also unable to supply the information, leaving the team without the material it needed.
For teams preparing PFMI self-assessments, responding to regulator requests for disclosure against specific principles, or mapping internal policies to the assessment methodology, this blind spot means the work cannot be completed accurately using AI tools alone. A team that does not recognise this limitation may proceed on the basis of generalised paraphrasing, submitting assessments that do not reflect the actual standard.
Regulatory submissions based on approximate rather than verbatim standards create a paper trail that is difficult to defend in supervisory review, and may require costly re-work or formal correction if the gap is identified by an overseer.