AI in action · Regulation

The SEC's 2027 Exam Priorities Call AI "Super Intelligence" and Tell Firms to Prepare for Deepfake Voices and Videos

The SEC's fiscal 2027 exam priorities, published October 23, 2026, list "emerging technologies, including super intelligence" among eight focus areas. Examiners will check how firms defend against AI-generated phishing and deepfake voice and video. For advisers that use algorithms, they will test whether the advice actually fits each client's stated profile. The Division of Examinations oversees more than 34,000 registrants.

Oct 24, 2026

Every autumn the Securities and Exchange Commission's Division of Examinations tells the firms it supervises what its examiners will look for in the coming year. The fiscal 2027 list, released on October 23, 2026 as press release 2026-109 with statements from Chairman Paul S. Atkins and Division Director Keith Cassidy, has eight focus areas for a population of more than 34,000 registered investment advisers, broker-dealers, exchanges, clearing agencies and other firms. The sixth is "emerging technologies, including super intelligence," a term the document uses throughout where earlier priorities said artificial intelligence; Wealth Management reported that the wording follows a presidential order on the term.

The description of how firms already use the technology is broad. "Registrants are increasingly incorporating super intelligence (SI) into their businesses," the document says, "using generative SI tools to support research, drafting, coding, and client communications, as well as advisory and portfolio management functions, and compliance monitoring and surveillance." Examiners will ask whether firms have policies covering that use, whether they supervise it, and whether what they tell clients about it is true.

Two passages turn the theme into specific tests. For automated advice and trading algorithms, exams will check whether "(1) representations are fair and accurate; (2) operations and controls are consistent with disclosures made to investors; and (3) algorithms lead to advice or recommendations consistent with investors' investment profiles or stated strategies." The third point is the one that matters to an investor: a model that recommends the same portfolio to a retiree and to a 25-year-old, whatever the questionnaire said, is a compliance failure, not a product feature. On security, the document warns that "threat actors are using SI to automate and accelerate cyberattacks, making those attacks more efficient and harder to detect," and says examiners will review how firms prepare for threats "such as SI-generated phishing messaging and synthetic or deepfake voice and video used for social engineering."

The same thread runs through the older rules. Identity-theft programmes under Regulation S-ID will be reviewed for whether they address "emerging risks, such as those related to SI and stolen credentials," and reviews of market infrastructure under Regulation SCI will cover super intelligence alongside tokenization, extended trading hours and quantum computing. The SEC also described its own use of the technology: its tools are "supplementing, but not replacing, staff judgment."

Exam priorities are not new rules, and the SEC's leadership has said it prefers to supervise AI with existing obligations rather than write a dedicated regime. But they are the closest thing firms get to the exam questions in advance. For a client, the practical reading is that a firm should be able to explain how its algorithm reached a recommendation, and how it would know that the voice on a call asking to move money was not its own executive.

Sources: SEC — SEC Announces Fiscal Year 2027 Examination Priorities, press release 2026-109 (Oct 23, 2026); SEC — Division of Examinations, Fiscal Year 2027 Examination Priorities (PDF, Oct 23, 2026); Wealth Management — SEC Sharpens AI Oversight Focus in 2027 Exam Priorities (Oct 23, 2026)