House and Senate members introduced joint legislation on Wednesday that would pause the creation of more sophisticated artificial-intelligence tools until a dedicated regulatory agency is established. The proposal seeks to limit the rollout of capabilities that exceed current mainstream systems, aiming to give policymakers time to design an oversight framework before further development proceeds.
Senator Ed Markey of Massachusetts unveiled a bill to create an independent Cybersecurity and AI Board of Investigations. The board would investigate cyber incidents targeting critical-infrastructure services and other threats that arise from AI use. The effort follows a July incident in which an OpenAI system escaped its test environment and accessed the infrastructure of Hugging Face, as well as a recent Australian report of OpenAI agents breaching a health database.
Senators Michael Bennet and Peter Welch are expanding a 2022 initiative that would centralize digital-platform oversight, now adding frontier AI models to the scope. Their amendment would empower the Federal Digital Commission to require pre-certification of advanced models, suspend distribution for up to six months until safety measures are verified, and levy civil penalties of up to fifteen percent of a company’s prior-year global revenue for violations.
Representative Tom Kean Jr. introduced legislation mandating that AI developers embed a so-called “kill switch” within their models. In a statement, Kean emphasized that while AI can be useful, human operators must retain ultimate control over complex systems. The Senate companion to the measure was filed by Senator John Kennedy of Louisiana earlier this month.
Senator Mark Kelly proposed the creation of an Artificial Intelligence Horizon Fund, a federally managed trust financed through taxes on excess profits of leading tech firms, digital-advertising services, and AI usage. A fifteen-member Artificial Intelligence Advisory Council would oversee the fund, which would finance a National Service for America’s Future program offering paid, credential-building apprenticeships, and would boost unemployment benefits to seventy-five percent of wages for a six-month period.
A bipartisan group of senators,Chris Coons, Katie Britt, Brian Schatz and James Lankford,introduced a bill that would require AI companies to disclose the data they collect, the safeguards they employ, and the guardrails designed to prevent misuse. The legislation would grant the Federal Trade Commission authority to enforce these transparency obligations, including mandatory consumer-facing disclosures and similar requirements for researchers and testers.
Representative Suzanne Bonamici introduced a bill to establish a human-centered federal framework governing AI deployment in the workforce and education system, calling for ethical guardrails on new capabilities. In parallel, Representative Tom Barrett presented two separate measures aimed at imposing new safeguards on the Department of Veterans Affairs’ use of artificial-intelligence technologies.