1. What did the SEC put on its 2027 examination list?
The Securities and Exchange Commission's Division of Examinations published its priorities for fiscal year 2027 on Thursday, October 23, 2026, and for the first time named artificial intelligence of the "super intelligence" variety as an area of heightened scrutiny. The eight-point list also places tokenization under market infrastructure risks, and it applies to more than 34,000 registrants.
The document describes "super intelligence and other advanced technologies that may affect market operations" as part of the agenda. That phrasing matters because it moves beyond the familiar question of whether a single firm uses an AI tool responsibly and asks how such systems might change the functioning of markets as a whole.
Annual priority lists are not rulemakings and they create no new legal duties. They do, however, tell the firms the agency supervises where examiners intend to spend their time, and registrants usually respond by reviewing their own controls in those areas before staff arrive.
2. How does the SEC treat AI as a market-wide variable?
In earlier editions of the examination priorities, artificial intelligence was largely discussed as a business tool that investment advisers and broker-dealers were beginning to adopt. The fiscal 2027 language frames the technology differently, as a factor that could influence how markets operate and not merely how one firm runs its back office.
The broader technology item in the list places AI under the heading of emerging technologies. One reasonable reading is that examiners will look at how AI feeds into compliance programs and into the detection of fraud, since those are areas where automated systems are already being deployed and where failures would be costly for investors.
For a firm, the practical implication is that statements made to clients about the use of AI, the supervision of automated processes and the records kept about how models are used can all become subjects of an examination. Firms that rely on third-party models also have to be able to show that they understand what those models do.
The Division also states that the list is not exhaustive. Examiners remain free to review any matter that touches compliance with federal securities laws across the entire registrant population, so the named priorities should be read as emphasis and not as a boundary.
3. Where does tokenization appear in the priorities?
Tokenization receives its own bullet under the heading of evolving market infrastructure. The priority area covers changes to that infrastructure, including reforms to Treasury clearing, tokenization and related developments, and it groups them as shifts in the plumbing through which securities are traded, cleared and settled.
Tokenization means recording ownership of a traditional asset, such as a share, a bond or a fund interest, as a digital token on a blockchain. Supporters argue that it can shorten settlement times and allow trading outside conventional hours. Supervisors, for their part, want to know how custody, record keeping and investor protection work when the record of ownership lives on a distributed ledger.
Broker-dealers, investment advisers and transfer agents that are building tokenized product lines should therefore expect examiner attention to follow their plans. The list signals that compliance programs for these products may be reviewed in practice and not only discussed in policy statements.
The pairing with Treasury clearing reforms is notable. Both items concern how market infrastructure is changing, and both involve firms that must demonstrate that new processes preserve the controls expected of older ones.
4. Who is covered by the program?
The examination program covers more than 34,000 registrants. They include investment advisers, investment companies, broker-dealers, clearing agencies and transfer agents, a population that spans small advisory practices and the largest firms in the industry.
Because the registrant base is so wide, the effect of a priority list differs by firm. A small adviser might see an examiner ask about how it supervises its use of generative AI tools for client communications, while a clearing agency might be asked about how it is preparing for tokenized collateral or changes to Treasury clearing.
The Division's director, Keith Cassidy, described the priorities as an adaptive approach. According to the SEC's release, the method lets the Division adapt to changing market conditions, identify patterns of risk and anticipate vulnerabilities for investors and for markets before they harden into losses.
That framing is consistent with how the examination function has long described its own work. The staff does not wait for a rule violation to be reported but samples firms, reviews documents and tests controls, and the priorities help determine where the sample is concentrated.
5. How does this compare with the previous year?
The fiscal 2026 priorities, published in November 2025, dropped crypto as a standalone focus area for the first time since 2021. At the time, the change was read as a sign that the SEC under Chair Paul Atkins was stepping back from the enforcement-heavy posture associated with the tenure of Gary Gensler.
The 2027 document does not bring a separate crypto section back. Instead, it folds the two technology themes that have shaped capital markets activity over the past year, artificial intelligence and tokenization, into the main list.
That structure suggests a shift in the way the agency talks about digital assets. Rather than treating them as a category apart, examiners are describing the technologies and the infrastructure changes that firms of every type are adopting, and applying the same compliance lens to each.
The change also fits with the description of a regulatory thaw under Atkins. A lighter approach to enforcement does not mean the absence of supervision, however, and the list shows that examiners still intend to look closely at how firms handle new technology.
6. What should firms prepare for?
Registrants that use or plan to use advanced AI should be able to document what the systems do, who is responsible for overseeing them and how the firm tests their outputs. Examiners typically ask for written policies, evidence that the policies are followed and records of how problems were handled.
Firms building tokenized products face a parallel set of questions. They should be ready to explain how ownership records are maintained, how customer assets are safeguarded, how trades are reconciled and how the firm would respond if the underlying technology failed.
Both areas call for coordination between compliance, technology and business teams. A control that exists only on paper, or a tool that was deployed before compliance was consulted, is the sort of gap that an examination tends to reveal.
Preparing before an examiner arrives is generally less costly than responding afterwards, since deficiencies found during an examination can lead to findings letters and, in more serious cases, referral to enforcement staff.
7. Why do examination priorities matter to markets?
The Division of Examinations is the arm of the SEC that inspects registered firms, as distinct from the enforcement staff that brings cases. Its work is routine and largely confidential, and most examinations end with a letter that identifies weaknesses for the firm to correct and not with a penalty.
Even so, the annual priorities are closely read because they shape where firms invest their compliance budgets. When a topic is named, chief compliance officers tend to receive questions from their boards, and consultants and law firms issue client alerts within days.
For investors, the list is a useful guide to the areas where the regulator sees risk building. Naming advanced AI and tokenization together indicates that the agency regards them as the two technologies most likely to alter how financial markets function in the near term.
The attention is also a reminder that the same technology can create benefits and hazards at once. Automated systems can spot fraud faster than people can, yet they can also concentrate errors, spread the same mistake across many firms using the same model, or make decisions that are difficult to explain afterwards.
8. What does systemic risk mean in this context?
Systemic risk refers to the possibility that a problem at one firm or in one technology spreads through the wider financial system. The decision to describe artificial intelligence as a variable that may affect market operations points to concern of that kind, as opposed to the narrower worry about a single firm misleading its clients.
If many advisers and trading firms were to depend on the same small group of advanced models, a flaw or outage in one of them could affect many participants simultaneously. Examiners are therefore likely to ask firms how dependent they are on outside providers and what they would do if a provider failed.
Tokenization raises a comparable question about infrastructure. A shared ledger can reduce the time and cost of settlement, but it also becomes a single point of reliance, so supervisors want to see that firms have tested contingency plans, access controls and recovery procedures.
None of these concerns amounts to a finding of wrongdoing. They describe the reasoning behind the priorities, and they explain why the agency groups technology adoption with the reform of market plumbing in a single list.
9. What happens next?
The priorities apply to the 2027 fiscal year, so examinations that reflect them will unfold over the coming months, and the agency typically publishes the next edition of the list about a year from now. Individual firm examinations are not announced, which means the first public sign of how the new emphasis plays out is likely to come from firms' own disclosures or from later statements by the Division.
Readers following the issue should watch for any additional guidance from the agency on artificial intelligence and tokenization, as well as for how firms describe their preparations. Until then, the list is best understood as a statement of where the staff intends to look, not a prediction of what it will find.