FINRA’s GenAI Playbook Promotes Real Accountability for Broker-Dealers

Financial Industry Regulatory Authority (FINRA) released its 2026 Annual Regulatory Oversight Report, showcasing a significant shift in focus towards generative artificial intelligence (GAI). Previously a topic of discussion, GAI is now a crucial part of daily operations for firms, as seen in the utilization of AI to draft communications, highlight suspicious activity, create policies, and streamline compliance. FINRA now expects GAI governance to meet the same standards as human employees, viewing algorithms as part of the supervision chain subject to examination.

The regulatory landscape has adapted to incorporate technological advancements like AI, with FINRA stressing that existing rules apply regardless of whether tasks are carried out by humans, traditional software, or AI tools. Specifically, utilizing GAI introduces unique risk factors that require increased scrutiny, especially when AI is utilized for regulated functions within firms. The integrity, reliability, and accuracy of AI models themselves need to be addressed in policies and procedures, emphasizing that attributing mistakes solely to AI will not serve as a valid defense.

To understand FINRA’s current focus, it’s essential to review its evolution regarding AI guidance. Over the years, FINRA has provided incremental guidance on AI, such as Regulatory Notice 24-09 and FAQs on advertising regulations concerning AI-generated content. However, the current report offers a newfound specificity, cataloging real-world use cases, observed practices, and emerging patterns in a manner well-prepared for regulatory examinations. This shift signifies a move towards accountability, with firms urged to view the report not as mere guidance but as a preview of expected queries and documentation during examinations.

FINRA notes a rapid increase in GAI adoption among member firms, primarily for internal efficiency improvements and information management. Some common applications include summarization, conversational AI, drafting documents, and workflow automation. While these applications offer efficiency gains, each introduces unique regulatory considerations that governance frameworks must address. The diversity in use cases underlines the reality that most firms are likely using GAI in some capacity, prompting the need for comprehensive governance frameworks to align with FINRA’s expectations.

Risk profiles associated with GAI are a significant focus for FINRA, notably issues related to accuracy, bias, privacy, cybersecurity, and the emergence of AI agents. Risks like hallucinations, bias, data drift, privacy breaches, and cybersecurity vulnerabilities are at the forefront of regulators’ concerns. Additionally, the autonomy of AI agents poses specific risks such as inadequate auditability, improper data handling, and misaligned incentives. Firms utilizing AI agents without human validation must establish stringent measures for oversight, transparency, and accountability to mitigate potential risks effectively.

In conclusion, FINRA’s revamped approach towards GAI governance signifies a critical shift towards ensuring accountability and regulatory compliance in firms leveraging AI technology. As the use of GAI becomes more widespread, broker-dealers must prioritize robust governance frameworks that address the unique risks associated with AI applications. The evolving regulatory landscape necessitates proactive measures to align with FINRA’s expectations and effectively navigate the challenges posed by algorithmic accountability.