Banks get AI testing pass—regulators may auto-approve risky experiments

What it does

This bill creates 'AI Innovation Labs' within financial regulators that allow banks, brokers, investment firms, and other regulated financial entities to test new AI-powered products and services under a temporary waiver from existing regulations. A company applies with an 'alternative compliance strategy' explaining how its AI experiment will still protect consumers and the financial system, and if approved, the regulator agrees not to enforce the normal rules for that specific test project—with a built-in end date and oversight requirements.

Why we flagged it

The bill's core mechanism is a regulatory waiver program—it suspends enforcement of existing financial rules for approved AI experiments. While framed as 'innovation,' the operative effect is to grant temporary immunity from regulation to financial firms testing AI systems, with consumer disclosure and risk management left largely to the applicant's own proposal.

What the text implies

• The 'deeming approved' clause (auto-approval after 240 days if the regulator does not act) creates a de facto approval mechanism that bypasses formal agency review—regulators may lack capacity to review all applications within the deadline, effectively forcing approval of marginal or risky proposals.
• The 'alternative compliance strategy' standard allows firms to propose their own risk-management approach in place of existing regulations; if the regulator approves it, the firm is shielded from enforcement for deviations from the original rule, even if the alternative fails to achieve the original rule's protective purpose.

The full analysis lists 5 implications of this text.

Who stands to gain

large financial institutions (banks, brokers, investment advisers); fintech companies and AI vendors serving financial services; securities firms and trading platforms