When Model Risk Management Meets Generative AI
Model risk management was designed for deterministic statistical models, and 2026 regulation is openly acknowledging that it does not fit generative AI: US SR 26-2 excludes it, the UK keeps SS1/23 technology-neutral, and the EU AI Act mandates tamper-evident logging from August. Validators now need replayable lineage and an independently verifiable record, not paperwork the bank produced about itself. The Open Audit Record, a subsystem of the Mickai SIOS, signs every action before it executes into a post-quantum hash-chained ledger a regulator can verify offline without trusting the institution.Originally published at https://mickai.co.uk/articles/model-risk-management-meets-generative-ai. Mickai is a Sovereign Intelligence Operating System; the Open Audit Record signs every artificial intelligence action before it executes, post-quantum and offline-verifiable. 101 filed UK patent applications, Mickai LTD (Companies House 17166618).
Authors
- Micky Irons (ORCID: https://orcid.org/0009-0000-5511-5858)
Institutions
- Maryland Institute College of Art (US)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-06-13
- DOI
- https://doi.org/10.5281/zenodo.20674211
- Primary Topic
- Ethics and Social Impacts of AI
- Type
- article
- Field-Weighted Citation Impact
- 0.00