0
Skip to content

AI Recordkeeping Poses Compliance Hurdles for Financial Services Firms

ai, tech, machine learningCompanies using artificial intelligence should keep existing record retention and supervision requirements in mind, even though regulators have yet to establish AI-specific rules.

09/17/2026 1:00 P.M.

Financial institutions are moving rapidly to integrate artificial intelligence (AI) across compliance, communications, and operational functions. However, adoption is raising a familiar regulatory challenge: how firms can prove compliance with rules written long before AI existed.

During a panel hosted by technology firm Red Oak, securities lawyers, compliance executives, and technology providers agreed that the lack of AI-specific regulations does not grant financial services firms a pass. Existing frameworks governing supervision, communications, recordkeeping, conflicts of interest, and fiduciary duties apply regardless of whether a human or an automated system performs the work, PYMNTS reported.

Enforcement actions are already occurring under current rules. Panelist Brian Rubin, a partner at Eversheds Sutherland and former Securities and Exchange Commission (SEC) enforcement attorney, highlighted recent SEC cases targeting “AI washing” — where companies exaggerate their AI capabilities.

Regulatory examiners expect firms to document who approved a tool, what data it accesses, how outputs are validated, and what human oversight is involved.

This approach creates complex recordkeeping questions, as no official rule dictates how long operational AI data must be stored, according to panelists. Rubin recommended anchoring strategies in existing standards. If an AI tool generates client communications, recommendations, or marketing materials, standard retention schedules apply.

Jamie Hoyle, vice president of product at MirrorWeb, cautioned firms against outsourcing accountability to vendors, noting “you’ll hear [vendors] talk about putting compliance on autopilot and you should run for the hills.”

The bottom line is that no matter the technology, regulated entities will still be held accountable. Documentation and human oversight remain the primary safeguards for managing AI compliance risks.

ACA’s Take

AI record retention is governed by existing legal and compliance frameworks rather than a standalone, industry-specific AI law. Regulators expect collection agencies to apply current consumer protection standards to all automated systems.

While Regulation F establishes a three-year federal baseline, many states impose longer retention periods for collection records. Several states explicitly mandate retention of “all communications” with consumers — a category that would logically include AI-assisted conversations.

Alaska, Hawaii, Idaho, Nevada, Oregon, West Virginia, and Wisconsin all require six-year retention periods. Collection agencies must honor any state retention requirements that are more restrictive than Regulation F.

This means a collection agency operating in multiple states may need to retain AI chat histories for six years or longer to comply with the most stringent state requirement.

Additionally, as Heath Morgan, partner at Martin Golden Lyons Watts Morgan, PLL, highlighted in an ACA Huddle, every AI interaction creates records — including prompts, outputs, timestamps, and user identifiers — that may be subject to litigation holds and discovery requests. Agencies that can’t identify where these records reside or who owns them face significant data exposure and spoliation risks.

Related Content from ACA:

 

Remember, subscribe to ACA Daily and Member Alerts under your My ACA Assistant profile when logged in to acainternational.org.

If you have executive leadership updates or other member news to share with ACA, contact our communications department at comm@acainternational.org. View our publications page for more information and our news submission guidelines here.

One moment please...

Share Profile

This site utilizes cookies as included in our Privacy Policy and User Agreement.