Financial firms must apply existing compliance standards to AI-generated communications, with FINRA maintaining that human judgement and verifiable records remain essential as artificial intelligence increases the volume of content requiring regulatory review.
RegTech provider Red Oak examined these challenges at FINRA’s Advertising Regulation Conference in Washington, D.C., on 15 and 16 October. AI governance was a recurring subject throughout the event, with discussions focusing on how firms supervise third-party AI tools, configure their behaviour and maintain records that can be reproduced for internal audits and regulatory examinations.
During the opening plenary, FINRA outlined its forward initiatives and regulatory developments, emphasising that its rules are technology-neutral. This means AI-generated communications must satisfy the same requirements as any other material, regardless of how they were created.
The regulator also highlighted the importance of human expert judgement during final reviews, given that large language models are not yet consistently accurate.
For compliance teams, this creates a growing review bottleneck. AI systems can generate content at a pace that exceeds the capacity of employees to assess it, while regulatory requirements remain unchanged. Human approval is still required, meaning the pressure point is increasingly the review process rather than content production.
The conference also examined the challenges of supervising social media activity and influencer promotions in financial services. A separate panel considered how firms monitor communications across different channels and maintain records of what has actually been published.
On off-channel communications, the panel stressed that a prohibition alone is not an effective control unless firms can also detect when employees are using those channels. Monitoring must extend beyond approved communication systems to identify activity that may otherwise go unseen.
Record retention presents another difficulty. Temporary and edited social media posts can be challenging to capture and preserve, particularly when platforms allow content to be changed after compliance approval. Panellists recommended comparing live material with the versions originally approved to establish whether published communications have been altered.
These issues place greater emphasis on firms’ ability to demonstrate how their AI systems operate, how communications are reviewed and whether the content ultimately published matches what received approval.
Red Oak’s AI Review tool is designed to support governance of third-party AI by using a firm’s own compliance handbook and applicable regulations as the basis for assessments. Teams can establish a working review process without writing prompts and decide what the system learns from reviewer decisions. The tool explains its findings and is intended to reduce review cycles as AI-generated content increases.
The company’s communications supervision agent, Mira, applies the same principles to monitoring. It reads a firm’s compliance handbook alongside SEC and FINRA regulations, explains its decisions in plain language, clears straightforward cases and escalates other matters to a human reviewer.
Every decision is timestamped, including those that do not reach a human reviewer. The resulting records can be exported when an examiner requests evidence explaining why a particular decision was made.
For supported platforms, Red Oak also captures edited, deleted and ephemeral messages. This gives firms the ability to compare live communications against approved versions and retain a record of what was visible at different points in time.
The approach is designed to keep human judgement central to compliance decisions. Firms define the policies governing the technology and retain responsibility for final decisions on escalated cases, while automation helps reduce routine reviews and focus attention on matters requiring human expertise.
As AI adoption increases, Red Oak’s approach highlights the importance of combining automated review with clear policies, human decision-making and records that firms can produce when regulators ask how their communications were supervised.
Copyright © 2026 FinTech Global
Copyright © 2026 RegTech Analyst
