SEC includes deepfakes in its 2027 examination priorities

The Securities and Exchange Commission (SEC), the US financial markets regulator, published its examination priorities for federal fiscal year 2027 on October 23, 2026. Among other things, it plans to examine firms’ preparedness for AI-generated phishing and fake voices or videos.

The SEC had already planned checks on AI-related cyber risks and oversight of AI use at firms in its 2026 program. The 2027 program specifies the threats targeted, citing AI-generated messages and synthetic voices or videos.

The SEC wants to examine how firms assess and prepare for these threats, according to the 2027 priorities report. A deceptive message or an impersonated voice may attempt to persuade a client or employee to disclose data or authorize a transaction.

The examination division covers more than 34,000 registered entities, including investment advisers, investment companies and brokers. This figure defines its remit, not the number of planned examinations. Federal fiscal year 2027 began on October 1, 2026, and will end on September 30, 2027.

Advice generated by algorithms

Examiners will also examine oversight of AI used in fraud detection, administrative operations, anti-money laundering and trading. The SEC also cites compliance with Regulation Best Interest, which governs brokers’ recommendations to retail investors.

The SEC plans to examine disclosures and controls associated with automated advisory tools and trading algorithms. In particular, it will look at whether algorithm-generated advice or recommendations match investors’ profiles or strategies.

These priorities guide risk-based examinations and may be adjusted during the year. Their publication does not create new legal obligations.