Democrats are preparing a slate of artificial intelligence bills for when they take control of Congress in January.
Rep. Bill Foster of Illinois got a head start Thursday (Nov. 5), releasing a discussion draft that would enlist the Federal Reserve in testing how regulators can spot financial stress as AI-powered services and round-the-clock payments accelerate its spread, according to a press release.
Foster’s Real-Time Regulatory Study Act of 2026 would direct the Federal Reserve Board to work with other financial regulators to develop, test and publicly release open-source demonstration software, the release said. The project would assess whether regulators could get real-time information about the health of banks and other financial institutions while making regulatory reporting easier and less costly. The Fed would seek technical contributions from financial institutions, technology providers, open-source organizations and researchers.
“With advances in AI creating new cybersecurity threats and financial risks, it’s more important than ever that we modernize the way we monitor the health of our financial institutions,” Foster said in the release.
Foster’s office said financial services powered by AI and payments that move at all hours can help transmit stress faster across the financial system, Roll Call reported Friday (Nov. 6). More timely information and automated tools could give supervisors a clearer picture of market conditions as they change. The tools could help institutions and regulators respond sooner, improve access to emergency funding and prepare for the orderly resolution of a failing bank.
The draft also addresses the cost of keeping regulators informed. A Foster spokesperson said in the report that the demonstration software would provide standard formats for regulatory data and tools for managing and transmitting it. That could reduce the manual work small institutions devote to reporting. Shared software could also lower development costs and give community banks and credit unions access to more advanced risk management tools.
Cybersecurity is part of the case for open sourcing the code. Foster’s office said that designing security into the software from the outset and allowing multiple contributors to maintain it could strengthen defenses as AI creates new threats, according to the report. The draft bill puts the Fed in charge of testing the software. Its immediate focus is on what the tools can do for supervisors and the institutions that supply their data.
Foster is the top Democrat on the House Financial Institutions Subcommittee and is in line to chair it next year, the report said. That would give him a platform to pursue the proposal as Democrats take over committee agendas. His draft adds a financial supervision initiative to a set of planned AI bills that range from model safety to the cost of powering data centers.
Rep. Lori Trahan of Massachusetts plans to pursue legislation requiring independent experts to test AI models against safety standards and clarifying when companies can be held liable for harm caused by their systems, Reuters reported Friday. Sen. Chuck Schumer of New York, who is expected to become majority leader, identified protecting consumers from higher electricity costs tied to data center expansion as a priority.
President Donald Trump’s opposition to mandatory AI rules could blunt the broader legislative push. But control of committees still gives Democrats room to hold hearings, compel testimony from industry executives and build support for specific proposals. In October, Democratic House committee leaders asked OpenAI, Anthropic and other developers to describe their security practices and report instances in which their AI tools acted in ways their designers had not intended, the report said.
For financial institutions, Foster’s draft puts a practical issue on that agenda. Supervisors need information fast enough to act when stress spreads, while community banks and credit unions must supply it with the staff and systems they have. The proposed Fed project would test whether common reporting formats and shared software can meet both needs.
For all PYMNTS AI coverage, subscribe to the daily AI newsletter.
