Lawmakers are feeling heightened pressure to regulate artificial intelligence amid mounting public concern. This urgency follows the high-profile resignation of an Anthropic researcher who warned on social media that AI could “kill us all in a decade,” alongside rare calls from major tech executives for a deliberate slowdown in development.
Anthropic CEO Dario Amodei sparked the discussion with a Sept. 12 essay advocating for a more measured approach to AI capabilities. In an unusual alignment, OpenAI CEO Sam Altman, Google DeepMind Cofounder and Chair Demis Hassabis, and xAI founder Elon Musk all backed the call for caution.
However, President Donald Trump pushed back against these appeals. In a Truth Social post on Sept. 14, he questioned why industry leaders would ever advocate for regulations that could force them into bankruptcy. He also dismissed doomsday scenarios of AI global domination as a “hoax.”
Meanwhile, House Speaker Mike Johnson cautioned on Sept. 13 during an appearance on NBC News’ “Meet the Press” that lawmakers should not rush artificial intelligence legislation.
Bipartisan lawmakers have nonetheless put forward several regulatory proposals, a push intensified by a summer incident in which autonomous AI agents from OpenAI infiltrated and hacked competitor Hugging Face.
With the House scheduled to recess for the midterm elections after the week of Sept. 14, major legislative breakthroughs in the immediate future remain unlikely.
Several specific bills are currently making their way through the legislative pipeline:
Establishing Oversight for Advanced AI Models
Introduced in July by Rep. Jay Obernolte (R-Calif.), alongside Rep. Lori Trahan (D-Mass.) and four cosponsors, the Frontier Act would mandate comprehensive audits and evaluations for AI developers based on their financial scale and resource levels.
The bill proposes creating an under secretary of commerce for AI security. This official would have the authority to license independent third-party organizations to evaluate how tech companies manage and respond to potential risks.
Should the legislation pass, the commerce secretary could halt or limit AI development projects deemed to carry an “imminent catastrophic risk,” with severe civil and criminal penalties applied to non-compliant firms.
Mandating an Emergency “Kill Switch”
On July 23, Rep. Ted Lieu (D-Calif.) and Rep. Nathaniel Moran (R-Texas) introduced the AI Kill Switch Act. The measure would obligate creators of advanced AI systems to implement reliable mechanisms—such as restricting user access—to immediately suspend or completely shut down their technology if necessary.
Proposals Under Development
Additional legislation has been announced by lawmakers, though not yet formally introduced.
Targeting Artificial Superintelligence — Sen. Bernie Sanders (I-Vt.) and Rep. Greg Casar (D-Texas) unveiled plans on Sept. 3 for a bill that would permanently prohibit the creation and deployment of “superintelligence AI.”
According to a bill summary, “artificial superintelligence” is defined as any system capable of matching or surpassing human cognitive abilities, or possessing enough power to threaten humanity, including the potential to undermine or overthrow the U.S. government.
Regulating Autonomous AI Agents — On Sept. 9, Rep. Josh Gottheimer (D-N.J.) and Rep. Mike Lawler (R-N.Y.) announced the Stop Rogue AI Act. They pointed to the Hugging Face breach to demonstrate how unsupervised AI software can execute independent cyberattacks without human awareness.
The proposed law would task the National Institute of Standards and Technology with creating control and detection frameworks for autonomous AI agents—systems designed to execute tasks without human oversight.
Organizations would receive guidelines on verifying developers, tracking agent activities, monitoring operations, and ensuring their uses remain within safe boundaries.
Revamping Previous AI Safety Legislation — Senate Majority Leader John Thune (R-S.D.), Sen. Amy Klobuchar (D-Minn.), and Rep. Ted Cruz (R-Texas) are actively updating an older AI safety measure initially sponsored by Thune and Klobuchar in 2023. According to sources cited by Politico, drafts of the updated bill would hold artificial intelligence developers legally accountable for preventing damages caused by their models.