Central banks are using artificial intelligence for faster, more precise data analysis, anomaly detection and risk management to improve efficiency, insights and decision-making. However, AI should never be a substitute for accountability, say the International Monetary Fund’s legal experts Marianne Bechara, Alessandro Gullo and Federica Giordano. Indeed, existing statutory immunities and good faith protections may be strained if officials rely on AI without meaningful oversight. Explainability should meet public-law transparency obligations. Therefore, internal safeguards and legal clarity on liability are essential, and legal teams are increasingly involved from AI procurement through deployment. From governance and audits to kill switches and exit rights, Bechara, Gullo and Giordano speak to Central Banking about how central banks can protect institutional autonomy.