Unleashing AI Innovation in Financial Services Act

This bill would set up “AI Innovation Labs” inside federal financial regulators. Banks, credit unions, and other regulated firms could test new AI tools under a clear plan approved by their regulator. During a test, some normal rules could be waived or adjusted if the firm follows an agreed “alternative compliance” plan that manages risks and serves the public, like improving access, security, or compliance, while avoiding systemic and national security risks and meeting anti–money laundering goals . Regulators must decide on applications within 120 days (with one 120‑day extension). If they still haven’t decided after the extension, the test is automatically approved. Regulators can still stop a test quickly if it threatens consumers, markets, deposit insurance funds, anti–money laundering rules, or national security, and they can always act against fraud or unsafe practices . Data given to agencies must be kept secure. Each agency has 180 days after enactment to set rules for how the labs work. Approved tests must last at least one year, and agencies must publish procedures for changes, penalties for breaking the plan, and coordination across agencies. Agencies must also send yearly, anonymous reports on results and lessons learned for several years .

Key points

• Who is affected: Regulated financial companies and their federal regulators; projects can involve multiple firms and more than one regulator if needed .
• What changes: Firms can test AI tools under an approved alternative compliance plan; during the test, only the terms of that plan apply for the specified rules, though fraud and unsafe practices remain fully enforceable .
• When: Agencies decide within 120 days (plus one extension); after that, approval is automatic. Agencies must issue program rules within 180 days. Each test runs at least one year, and annual reports begin within two years of enactment and continue for seven years .