Regulators Can't Crypto-Splain AI Yet, So They're Supervising the Humans Instead 🤖

Financial supervisors are applying existing AI oversight frameworks to crypto firms, even as no dedicated crypto-specific AI rule has been issued, according to regulatory comments gathered in recent days. The supervisory focus centers on human intervention in AI systems, rigorous testing, data quality, and lifecycle monitoring, with direct implications for exchanges, stablecoin issuers, and blockchain analytics providers that use machine learning for transaction monitoring, sanctions screening, and risk decisions.

Bank of England Governor Andrew Bailey, writing in a September 30 article, emphasized the need for meaningful human intervention in AI systems and highlighted the importance of rigorous testing to understand risks, including cybersecurity threats in finance. European Central Bank Supervisory Board member Patrick Montagner, speaking on October 1, stressed the need for clearly defined responsibilities, high-quality data, and lifecycle monitoring for banks using AI. Fernando Restoy of the Bank for International Settlements added that AI is changing banking operations, requiring supervisors to address broader financial resilience alongside individual-firm controls.

The supervisory push arrives as stablecoin-settled transactions expand, agentic payment systems develop, and criminal actors adopt more sophisticated tooling. Regulators have not introduced a single crypto-specific AI rule, leaving firms to operate under general principles that emphasize explainability, auditability, and human oversight of automated risk decisions affecting $BTC, $ETH, and other digital asset markets.

Industry participants say the open question is no longer only whether AI can accelerate transaction monitoring, but whether firms can explain, audit, and supervise the systems that inform their risk decisions. Compliance teams at major exchanges and stablecoin issuers are responding by documenting model governance, tightening change-management procedures, and expanding human review queues for AI-flagged transactions, according to public statements and supervisory guidance over the past quarter.

Supervisors have signaled that examinations will test whether crypto firms maintain effective human oversight of AI systems used for compliance and risk management, regardless of whether a dedicated crypto AI rule is in place.