AI Regulation Tracker  /  Guidance and standards

SEC Flags AI as a Cross-Cutting Focus in Its 2026 Examination Priorities

On November 17, 2025, the SEC Division of Examinations released its 2026 examination priorities. Artificial intelligence runs through them. The Division says it will check whether a firm's claims about its AI match what the firm actually does, and whether the firm supervises its AI use, including for anti-money laundering. These are priorities, not a new rule, but they tell registered advisers and broker-dealers exactly where examiners will look.

By Anthony Guerriero  ·  The Leveraged Years Regulatory Research  ·  Published July 20, 2026  ·  Last updated July 20, 2026

Once a year the SEC Division of Examinations publishes a document that tells regulated firms where its examiners plan to spend their time. The Division is blunt about why it does this. As it puts it, the priorities exist "to provide transparency to registrants and investors about the topics that the Division plans to focus on in the new fiscal year and to encourage firms to direct their compliance efforts on areas of potentially heightened risk." SEC Chairman Paul S. Atkins framed the same point in the release, saying examinations "should not be a 'gotcha' exercise" and that the priorities should let firms prepare for a constructive conversation with examiners.

So this is not an enforcement action and it is not a new rule. It is a map. For fiscal year 2026, released on November 17, 2025, that map puts artificial intelligence squarely in the frame, and it does so in language that finance and compliance professionals should read closely.

Where AI actually appears in the priorities

The clearest AI passage sits in the Division's section on emerging financial technology. The Division says it remains focused on registrants' use of tools such as automated investment tools, AI technologies, and trading algorithms. Then it gets specific about what examiners will test. In its own words:

"With respect to AI, the Division will focus on recent advancements in AI and will review for accuracy registrant representations regarding their AI capabilities or AI."

Read that twice. The Division is telling firms it will check whether the claims they make about AI are accurate. If a firm tells clients or regulators that AI drives its portfolio decisions, examiners will look for evidence that the AI genuinely does that, rather than sitting in a slide deck. This is the risk the market has taken to calling AI-washing, and the exam standard for it is simple: do your representations match your practice.

The next sentence widens the lens to supervision and controls. The Division says:

"The Division will assess whether firms have implemented adequate policies and procedures to monitor and/or supervise their use of AI technologies, including for tasks related to fraud prevention and detection, back-office operations, anti-money laundering (AML), and trading functions, as applicable."

That single sentence carries a lot. It ties AI directly to written policies and procedures, to supervision, and to anti-money laundering. It means a firm cannot simply deploy an AI tool in AML monitoring or trading and leave it unsupervised. The Division expects a governance layer around it: documented policies, a way to monitor the tool, and a way to supervise how staff use it.

AI in cybersecurity and AML

AI is not confined to the technology section. In its cybersecurity coverage, the Division says its focus includes the training and security controls firms use to identify and mitigate new risks associated with, in its words, "artificial intelligence (AI) and polymorphic malware attacks." In other words, examiners are thinking about AI as both something firms adopt and something adversaries use against them.

And anti-money laundering gets its own section, grounded in the Bank Secrecy Act, which requires broker-dealers and certain registered investment companies to maintain AML programs reasonably designed to prevent money laundering and terrorist financing. Pair that binding obligation with the earlier line about supervising AI in AML tasks, and the message is coherent: if you use AI inside your AML program, the AML program still has to work, and you still have to be able to show it works.

What this is, and what it is not

I want to be precise, because the enforcement mechanics matter and it is easy to overstate a priorities letter.

The 2026 examination priorities are not a regulation. They do not create a duty that did not exist on November 16, 2025. The Division says as much: the priorities are not an exhaustive list, and the scope of any exam also turns on a firm's history, operations, and products. Nobody is newly liable because AI was named a priority.

What is binding is the body of law the priorities measure against. An investment adviser owes a fiduciary duty. Advisers must adopt and follow written compliance policies and procedures under the compliance program rule. Broker-dealers and certain funds must maintain AML programs under the Bank Secrecy Act. Representations to clients must be accurate under the antifraud provisions of the securities laws. The priorities do not add to those duties. They tell you that AI is now one of the lenses examiners will use to test whether you are meeting them. The map is not the law, but it tells you exactly where the enforceable lines already run when AI is in your business.

What this means for advisers and broker-dealers

If you are an SEC-registered adviser or a broker-dealer, treat the AI passages as an exam prep checklist rather than an alarm. Three things carry the weight.

First, reconcile your AI claims with your AI practice. Walk through your Form ADV, your marketing, your website, and your client communications, and find every place you describe AI. For each one, ask whether you can show, with evidence, that the AI does what you say it does. If it cannot, fix the representation or fix the practice before an examiner does it for you.

Second, put governance around any AI you actually use. The Division is looking for policies and procedures that monitor and supervise AI, especially in fraud detection, back-office work, AML, and trading. That means written procedures, a named owner, a way to test the tool, vendor oversight where the model is third-party, and a record that a human is supervising the output.

Third, document it now. The through-line in every AI passage is evidence. Examiners will ask what your AI does, how you supervise it, and how you know your disclosures are accurate. The firms that struggle will be the ones that used AI enthusiastically and wrote none of it down. The firms that do well will have a governance file ready before the exam letter arrives.

Browse the full AI Regulation News tracker

Informational analysis for working professionals, not legal, compliance, or investment advice. Confirm how the SEC examination priorities and the underlying securities laws apply to your firm with qualified counsel and your compliance team.