Society GUIDE
SEC and FINRA Rules on AI for Advisors
No SEC or FINRA rule is written specifically for AI.
On this page4 min read
Overview
Broker-dealers and registered investment advisers must apply their existing supervision, communications, recordkeeping, best-interest and fiduciary rules to AI tools just as they would to any other technology. FINRA said this directly in Regulatory Notice 24-09 in June 2024. It matters because a firm stays fully responsible for what a chatbot says, what a note-taker records and what a vendor's model recommends.
Deep Dive
Both regulators describe their rules as technology-neutral: what matters is the activity, not the tool that performs it.
For FINRA member broker-dealers, the key rules are these. Rule 3110 requires a supervisory system reasonably designed to achieve compliance, which covers how AI tools are approved, monitored and tested. Rule 2210 governs communications with the public. A written communication distributed or made available to more than 25 retail investors within any 30-calendar-day period is a retail communication. One sent to 25 or fewer is correspondence. Each category has its own content standards and review requirements, and firms have to decide how chatbot output and AI-drafted material fits. Rule 4511, together with Securities Exchange Act Rules 17a-3 and 17a-4, requires firms to create and keep business records. That can include prompts and outputs when they are business communications. Regulation Best Interest applies whenever a recommendation reaches a retail customer, whether a person or a model produced it.
For registered investment advisers, the Advisers Act fiduciary duty of care and loyalty applies. So do the compliance program rule (Rule 206(4)-7), the books and records rule (Rule 204-2) and the Marketing Rule (Rule 206(4)-1), which covers any claims a firm makes about its own AI.
In July 2023 the SEC proposed a rule on conflicts of interest from predictive data analytics. The Commission withdrew that proposal in June 2025. A common misconception is that the withdrawal left AI unregulated. Conflicts, supervision failures and misleading statements are still addressable under existing rules, and SEC examination priorities have continued to mention AI.
Vendor tools raise their own issue. FINRA has said that outsourcing a function does not outsource the responsibility for it.
Strategic Impact
Risk and safety
Catastrophic and everyday AI harms both depend on who understands the risks and who can act.
Clearer decisions
Public and professional literacy shapes whether strong safety policy is politically possible.
Cutting through hype
Clear explanations reduce capture by hype, lab PR, and vague ethics theater.
The Future of SEC and FINRA Rules on AI for Advisors
Regulators have so far relied on existing rules, guidance, exam findings and enforcement rather than AI-specific rulemaking, and the 2025 withdrawal of the predictive analytics proposal fits that pattern. Firms should expect more detailed exam questions about AI inventories, vendor oversight and record capture. They should also expect enforcement based on familiar theories such as misleading statements or inadequate supervision. FINRA has asked members for input on how its rules apply to newer tools, including agent-style systems, so further guidance is plausible. The durable approach is to design controls around each existing obligation instead of waiting for AI-specific rules.
Real-World Implementation
A broker-dealer uses a generative AI tool to draft social media posts. Each post is treated as a retail communication under FINRA Rule 2210: it must be fair and balanced, and it needs principal approval where the rule requires it.
An RIA's AI note-taker produces meeting summaries that are emailed to clients. The firm keeps those summaries as books and records and reviews samples to confirm they reflect what was actually discussed.
A firm licenses a vendor's model that suggests portfolio changes. Before deployment, compliance does due diligence on how the model works, what data it uses and how its outputs are tested, because outsourcing does not transfer the firm's obligations.
A firm tests a client-facing chatbot and blocks it from making individualized security recommendations. Under Regulation Best Interest, those recommendations would have to be in the retail customer's best interest and supervised accordingly.
Risks & Guardrails
Treating existential risk as sci-fi while capability compounds.
Confusing surface product safety with alignment under high autonomy.
Leaving non-English and non-expert audiences with only low-quality sources.
Implementation Roadmap
Separate product harms, misuse, and loss-of-control / misalignment risks.
Ask what evidence would change your view on timelines and severity.
Prefer primary sources and concrete evals over marketing claims.
Identify one action path: career, policy, funding, or skills — not only awareness.
Keep Exploring
Free newsletter
Keep up with AI in 3 minutes a day
One short email each weekday with the three AI stories that actually matter. Free forever, no ads.
One email each weekday. Unsubscribe in one click. We never sell or share your address.
Test yourself
Take the SEC and FINRA Rules on AI for Advisors quiz
Instant feedback on every answer, and a shareable certificate with a verifiable ID once you pass a course.
Support free AI education. AI Understanding is a 501(c)(3) nonprofit — no ads, no paywall, ever. Make a donation
Frequently asked questions
What is SEC and FINRA Rules on AI for Advisors?
No SEC or FINRA rule is written specifically for AI. Broker-dealers and registered investment advisers must apply their existing supervision, communications, recordkeeping, best-interest and fiduciary rules to AI tools just as they would to any other technology. FINRA said this directly in Regulatory Notice 24-09 in June 2024. It matters because a firm stays fully responsible for what a chatbot says, what a note-taker records and what a vendor's model recommends.
What was the central message of FINRA Regulatory Notice 24-09 on generative AI?
The notice reminded members that their existing obligations apply whatever technology they use, including generative AI.
Which FINRA rule requires a broker-dealer to have a supervisory system reasonably designed to achieve compliance, including for AI tools?
Rule 3110 is FINRA's supervision rule. Rule 2210 covers communications, and Rule 4511 covers books and records.
Under Rule 2210, when does an AI-drafted written message become a retail communication rather than correspondence?
The dividing line is audience size over a 30-day period: more than 25 retail investors makes it a retail communication.
What happened to the SEC's July 2023 proposal on conflicts from predictive data analytics?
The SEC withdrew the proposal in June 2025. Existing rules still cover conflicts and misleading statements.
An RIA advertises that its portfolios are 'AI-optimized.' Which rule most directly governs whether that claim is allowed?
The Marketing Rule governs advisers' advertisements, including claims about their own AI.
Keep learning
Related guides
More guides picked for this topic