Advancing FINRA’s Mission With AI
By Robert Cook, President and CEO, FINRA
FINRA’s member firms have long been at the forefront of utilizing innovative technologies to better serve their clients and support vibrant capital markets. Indeed, their capacity to foster and embrace advancements in technology is a key reason why U.S. securities markets are widely considered the envy of the world.
Like our member firms, FINRA also has a long track record of innovating with technology. For example, in the 1970s our predecessor, the NASD, developed a nationwide clearing system for the over-the-counter (OTC) market and established Nasdaq to automate trading in OTC securities. A decade ago, FINRA was an early adopter of cloud computing and storage to more efficiently ingest and analyze very large data sets arising from our regulatory programs. And today, FINRA’s technology has evolved to support our oversight of more than 3,200 member firms, 630,000 registered persons, and their sales and trading activities in the equities, options, and fixed income markets.
Now as part of our FINRA Forward initiative, FINRA has embarked on the latest phase in its technology journey—building expertise, infrastructure and capabilities in GenAI to perform our self-regulatory functions more effectively and efficiently.
In this and other areas of our work, we benefit greatly from ongoing engagement with our member firms as part of our self-regulatory model. There is a constructive feedback loop between our internal efforts to adopt GenAI and our interactions with member firms to understand their adoption of GenAI—what we learn in one context enriches our work in the other.
In this spirit, I want to share how FINRA has begun using GenAI to support our regulatory program, as well as how we have been engaging with our member firms regarding their use of GenAI.
FINRA’s Early Adoption of AI
FINRA’s adoption of AI began some time ago with our market surveillance programs. FINRA has an obligation to oversee the market activity of our member firms wherever they trade, which today could be across dozens of exchanges, alternative trading systems, and other venues. And pursuant to contractual arrangements, we perform cross market surveillance for 26 SROs operating 35 equities and options exchanges. As securities markets grew increasingly complex and dispersed and volumes soared, it became impossible to rely on traditional spreadsheets or similar tools to analyze the relevant data. Thus, for years FINRA has by necessity been developing and employing AI to support its market oversight functions.
Today, FINRA runs many algorithms that use AI to review the hundreds of billions of market events generated each day in search of fraud, manipulation, or other misconduct that can harm investors and markets. When this surveillance indicates potentially problematic activity, we investigate further or, for matters outside our jurisdiction, we make a referral to the SEC or another agency.
Importance of a Strong Governance Framework for Approaching AI
Our extensive experience with using AI in our market surveillance work has informed our more recent efforts to leverage large language models (LLMs) and GenAI-enabled tools in FINRA’s broader regulatory program.
Given our regulatory role and responsibilities, we recognize the need to balance innovation and experimentation with having effective controls in place before implementing any new AI tool. To that end, FINRA approaches new AI-enabled tools carefully, to allow time for piloting and refining the technology to help ensure that the tool meets its intended purpose and that potential risks are identified and addressed. Our overall process is coordinated by an internal AI Governance Committee. Recognizing the importance of data privacy, we limit access to sensitive data within FINRA based on regulatory need and security considerations. And we maintain comprehensive oversight of AI-enabled tools through rigorous security assessments and regular reviews to ensure the tools continue to meet our security standards and safeguard regulatory information.
FINRA’s effectiveness in serving our mission depends on our people, and we want to encourage and empower them to use AI thoughtfully and effectively. We provide training to build expertise, share different use cases, and explain how AI can and cannot be used at FINRA.
This year, FINRA made foundational GenAI training available to all staff, along with more specialized training on different tools and how to use them. FINRA also has offered an immersive learning program to build employees’ expertise in using data analytics to inform decision-making and solve complex problems. To date, more than 600 FINRA staff have completed the program and more than 900 FINRA staff have completed at least one GenAI-focused training.
New GenAI-Enabled Tools to Advance FINRA’s Mission
Guided by this internal framework, FINRA is identifying and implementing opportunities to use GenAI to strengthen our regulatory program by more quickly generating valuable insights from data, performing or streamlining certain work processes, and freeing up employees’ time for higher value uses.
Last year FINRA introduced an internal LLM-based chat capability, called FILLIP, that employees have leveraged to support a range of functions, such as summarizing and analyzing information, comparing documents for material changes, and writing and editing drafts. FINRA staff also have used FILLIP to assist in a variety of regulatory tasks, such as conducting member firm risk reviews and analyzing data on mutual funds and ETFs to facilitate examinations of related sales activities. FINRA started piloting FILLIP with select departments in the fall of 2024 and rolled it out to all FINRA staff in May 2025. About 40 percent of FINRA employees already use FILLIP each week. FILLIP ambassador-mentors and regular learning labs are increasing adoption and highlighting use cases and best practices.
We also have developed dedicated GenAI-enabled tools to facilitate more specific work processes. These have included a tool to summarize and analyze investor complaints filed directly with FINRA (which in recent years have exceeded more than 10,000 per year) for faster review and response by investigators, and a tool to summarize and analyze comments on our Regulatory Notices to aid us in more quickly understanding and considering feedback on rule proposals. FINRA has other GenAI-enabled tools in the internal development or experimentation phases, including a tool to summarize and evaluate disclosures filed with FINRA’s Credentialing, Registration, Education and Disclosure (CRED) department and a tool to extract and compare information from member firms’ eFOCUS reports.
FINRA also leverages vendor-created GenAI tools in areas such as assisting in software development and maintenance, creating learning content, and providing internal technical support to augment our help desk capabilities and better serve staff needs. In addition, we encourage our staff to identify potential new use cases. For example, at FINRA’s 10th annual Createathon in June 2025, several internal teams built prototype GenAI-enabled tools for increasing our regulatory effectiveness.
These examples are representative of the different uses for GenAI we are pursuing. Collectively, we are estimating many thousands of hours in annual efficiency gains from staff using the GenAI tools that are already deployed or in development, with more to come. While achieving these efficiencies is exciting, our staff’s expertise and knowledge is and will remain crucial to successfully leveraging GenAI, especially when maintaining a “human-in-the loop” as part of our governance and controls.
Engaging with Member Firms on GenAI
Like FINRA, member firms are exploring the new opportunities and risks presented by recent advancements in GenAI.
FINRA’s rules—and many securities law requirements more generally—are intended to be technology neutral and continue to apply when member firms use AI. FINRA has previously identified certain regulatory considerations and provided resources for member firms when using GenAI. But as the technology evolves and member firms continue to consider different GenAI use cases, we understand additional guidance may become appropriate, and FINRA stands ready to work with firms on potential supervisory and compliance implications. Regulatory Notice 24-09 (June 2024) has additional information on opportunities for members to engage with FINRA on these matters.
As noted above, FINRA benefits from ongoing interactions with our member firms, and we have the opportunity to share with them what we learn from our oversight activities. Consistent with this framework, we are engaging with member firms to understand and support their use of GenAI, including through regular Risk Monitoring outreach, discussions with our advisory committees, industry roundtables, and dedicated conference sessions and events. To help further inform member firms, we recently published a summary of the most common GenAI use cases that FINRA has observed among member firms.
FINRA also has kept member firms informed of the risks arising from the potential adversarial use of GenAI by bad actors, including through a continuing series of Threat Intelligence Products (TIPs) that are delivered directly to key member firm personnel. Recent TIPs have advised firms about the ways in which bad actors might use GenAI to enable and enhance various scams and frauds, including investment club scams, new account fraud and account takeovers, and business email compromise scams. We also are planning a webinar for early next year to help member firms protect themselves and their investors from the adversarial use of GenAI.
Moving forward, FINRA will continue to reinforce this constructive feedback loop by learning from our member firms about their experiences with GenAI and in turn sharing with them what we have learned through our own experiences and our oversight activities. This feedback loop ultimately enables FINRA and our member firms to better utilize AI in a manner that serves the interests of investors and promotes innovative and vibrant capital markets.