FSB consults on responsible AI adoption in finance
The Financial Stability Board (FSB) has opened a consultation on sound practices for the responsible adoption of artificial intelligence (AI) by financial institutions, published on 10 June 2026. Comments are invited until 22 July.
The report sets out 12 sound practices spanning organisation-wide governance and the full AI lifecycle, from inception through to retirement. It covers everything from board-level oversight and risk appetite to data governance, explainability, human oversight and the particular challenges posed by generative and agentic AI.
The FSB last assessed the financial stability implications of AI in 2024, but the technology’s rapid evolution and accelerating adoption across the financial industries has prompted this more practice-focused follow-up.
The recommendation for Sound Practice 7, on data governance, says institutions should keep their data accurate, complete, consistent, reliable and secure across training, testing and use. That’s familiar territory for anyone working with structured data: an AI model is only ever as good as the data you feed it, and messy or unreliable inputs don’t magically become trustworthy outputs because there’s a clever model in between.
Read more and respond to the consultation here.
