A substantive restructuring and expansion. v2 renumbers sections, redefines "AI test project" and "financial product, service, or activity," and significantly expands the application requirements and agency review process for AI Innovation Labs. It also adds new provisions for interagency coordination and a notice of intervening agency.
Unleashing AI Innovation in Financial Services Act
- Sponsor
- Rep. Hill, J. French [R-AR-2]
- Committees
- Financial Services Committee (primary)
- Last action
- Jun 24, 2026
Bottom line
The bill creates a regulatory sandbox within federal financial agencies to accelerate AI innovation in financial services by offering conditional regulatory relief and oversight for approved experimental projects.
What it actually does
This bill mandates that each financial regulatory agency establish an "AI Innovation Lab" to allow regulated entities to experiment with artificial intelligence (AI) in financial products, services, or activities. Participants can apply for waivers or modifications of existing regulations, provided they propose an alternative compliance strategy and demonstrate public interest benefits and robust risk management. The labs aim to foster innovation by reducing regulatory burdens and enforcement risks for approved AI test projects.
Proponents argue
Proponents argue that the bill is essential for fostering responsible innovation in the rapidly evolving field of artificial intelligence within the financial sector. They contend that by providing a controlled environment for experimentation, the bill will enable regulated entities to develop new AI-powered financial products and services, enhance efficiency, improve risk management, and strengthen cybersecurity, ultimately benefiting consumers, investors, and the competitiveness of the U.S. financial system.
Opponents contend
Opponents might express concerns that creating regulatory sandboxes could inadvertently lead to an erosion of consumer and investor protections, or create new systemic risks if AI test projects are not rigorously overseen. They may argue that the complexity of AI technologies could make it difficult for regulators to adequately assess and mitigate novel risks, potentially creating loopholes that could be exploited or lead to unintended negative consequences for financial stability.
The bill is relatively concise and clearly structured, allowing for a thorough review within typical legislative review periods.
Section 3
Establishment of AI Innovation Labs
This section mandates that each federal financial regulatory agency, such as the SEC, FDIC, and Federal Reserve, must establish or designate an "AI Innovation Lab." The purpose of these labs is to create an environment where regulated financial entities can experiment with AI test projects without facing unnecessary regulatory burdens or enforcement actions, provided their projects are approved.
Supporters argue
Proponents argue that establishing dedicated innovation labs is crucial for agencies to proactively engage with AI development, providing a structured environment for experimentation that balances innovation with necessary oversight. This approach helps agencies build expertise and adapt regulations to new technologies.
Critics contend
No specific opposition is typically raised against the establishment of innovation labs themselves, though concerns may arise regarding their funding, staffing, and the scope of regulatory relief they offer.
Tradeoffs
The provision balances the need for regulatory flexibility to foster innovation with the imperative for agencies to maintain oversight and protect financial stability. It requires agencies to invest resources in understanding and managing new technologies.
Section 4(a)(2)
Application Requirements for AI Test Projects
This section details the comprehensive information regulated entities must submit when applying to engage in an AI test project. Applications must describe the AI technology, propose an alternative compliance strategy for specific regulations, explain how the project serves the public interest and manages risks (including cybersecurity), specify a termination date, and outline a business plan and estimated impacts.
Section 4(a)(7)(B)
Regulatory Relief and Enforcement Limitations
Upon approval of an AI test project, the appropriate financial regulatory agency may only enforce the waived or modified regulation in the manner specified in the approved alternative compliance strategy. This means that for the duration of the test project, the regulated entity is protected from standard enforcement actions related to that specific regulation, as long as it adheres to its approved alternative plan.
Section 4(a)(4) and (6)
Joint Agency Review and Intervention
If an AI test project involves regulations from multiple financial regulatory agencies, the application must be submitted to all relevant agencies and requires joint approval. Furthermore, any financial regulatory agency that receives notice of an application and has jurisdiction over the regulation in question can issue a 'notice of intervention,' which then also requires their joint approval for the project to proceed.
Section 4(a)(7)(D)
Deemed Approval for Untimely Agency Decisions
If a financial regulatory agency fails to approve or deny a complete application for an AI test project within the initial 120-day review period (or an extended 240-day period), the application is automatically deemed approved. This provision places a strong incentive on agencies to act promptly on applications.
Section 4(a)(9) and (b)
Mandatory Agency Rulemaking and Coordination
The bill requires each financial regulatory agency to promulgate implementing regulations within 180 days, covering application procedures, consequences for non-compliance, termination dates, reporting, disclosures, and confidentiality. Additionally, agencies are mandated to consult, share information, and coordinate to establish best practices, update guidance, and avoid duplicative efforts.
Section 5
Annual Reporting on Outcomes
Each financial regulatory agency is required to submit an annual report to Congress for seven years, detailing the aggregated findings, trends, and lessons learned from the AI test projects. These reports must not include proprietary or confidential business information and may contain a classified annex where appropriate.
If a financial regulatory agency fails to approve or deny a complete application for an AI test project within 120 days (or 240 days with an extension), the application is automatically deemed approved.
Section 4(a)(7)(D)
Why it matters:While standard legislative practice often places procedural details within larger sections, the 'deemed approval' clause is a significant policy lever that could be easily overlooked. Its placement might be seen as a way to ensure agencies are pressured to act without making the mechanism overly prominent.
Case for: Supporters would argue this clause is a necessary safeguard against bureaucratic delays, ensuring that innovative projects are not stifled by regulatory inertia. It places accountability on agencies to be efficient.
Case against: Critics might contend that automatic approval, even after a set period, could lead to the approval of projects that have not received adequate scrutiny, especially if agencies are under-resourced or overwhelmed. This could compromise risk management and consumer protection.
Estimated impact: Could significantly impact the speed of AI project approvals and the operational burden on financial regulatory agencies.
The definition of 'financial product, service, or activity' includes 'any internal activity of a regulated entity that is carried out in furtherance of the offering or sale of a product, service, or activity described in subparagraph (A).'
Section 2(6)(B)
Why it matters:This is standard legislative practice for defining scope. It's not necessarily evasive, but its significance for the breadth of the sandbox could be missed.
Case for: Proponents would argue that including internal activities is crucial because AI can significantly enhance back-office operations, risk management, and compliance processes, which indirectly benefit consumers and the financial system. Limiting the sandbox to only external products would miss a large area of potential innovation.
Case against: No specific opposition is typically raised against this broad definition, as it aligns with the goal of comprehensive AI innovation. However, some might argue that internal activities, being less visible, could pose different oversight challenges.
Estimated impact: Significantly expands the potential scope of AI test projects beyond direct consumer offerings to include internal operational efficiencies and risk management tools.
Nothing in the provisions granting regulatory relief 'may be construed to limit the authority of a financial regulatory agency to take an enforcement action against a regulated entity with respect to fraud or market manipulation or for engaging in an unsafe or unsound practice.'
Section 4(a)(7)(F)
Why it matters:This is a standard legal disclaimer to ensure that core regulatory powers related to market integrity and safety are not inadvertently curtailed by new legislation. Its placement is typical for such saving clauses.
Case for: Proponents would emphasize that this clause is vital for maintaining the integrity of financial markets and protecting consumers. It ensures that the sandbox does not become a shield for malicious or reckless behavior, reinforcing the principle of responsible innovation.
Case against: No specific opposition is typically raised against this type of anti-fraud or market manipulation saving clause, as it is a fundamental aspect of financial regulation.
Estimated impact: Maintains critical enforcement powers for financial regulatory agencies, ensuring that the regulatory sandbox cannot be used to circumvent fundamental protections against fraud, manipulation, or unsafe practices.