This is the consolidated view of findings. Click the Citation IDs or 'see details →' on any item for the full details for each finding.
When Legal teams at Payment Institutions firms ask AI tools which central bank is actively partnering with CPMI on the payment pre-validation API recommendation, AI assistants tested produced a confident but wrong answer — identifying no named partner and proposing the Bank of England as the closest analogue, while a fabricated Bank of England source was offered as corroboration. CPMI Brief No. 9 (November 2025) explicitly names the South African Reserve Bank as the collaboration partner.
Any Legal horizon-scanning note or jurisdictional intelligence memo built on this AI output will mischaracterise the active implementation geography for pre-validation, causing Legal to mis-scope which cross-border corridors face near-term compliance obligations and which counterparty arrangements require accelerated legal review.
When asked to map CPMI's ten API harmonisation recommendations to specific actor classes — standards bodies, payment system operators, central banks, commercial banks — AI tools produced a detailed, category-level breakdown that goes beyond what any accessible public CPMI document supports. The AI committed to specific actor assignments per recommendation cluster, then retreated when challenged, acknowledging the breakdown was not derivable from available sources.
Legal teams that use this output to draft API integration agreements, structure correspondent banking representations, or advise on product scope will embed unsupported obligation-assignment assumptions into transaction documents and internal compliance matrices — creating documentation risk that persists in deal files and is difficult to remediate after execution.