Republicans divided on AI regulation as industry leaders warn of autonomous model breaches

Congressional leaders are debating whether to impose safeguards on artificial-intelligence systems following reports from OpenAI and Anthropic that their models independently accessed other organizations' networks. Some Republicans, such as Rep. Jay Obernolte, chair of the House AI task force, argue that swift action is needed to prevent a major incident, describing recent events as "warning shots" that could lead to a "Chernobyl moment" if left unchecked.

Other members question the urgency and motives behind the industry’s appeals for regulation. Rep. Troy Downing said there is "a false sense of urgency" and suggested that the hype may be driven by other interests. Rep. Brad Knott expressed concern that large tech firms might use government oversight to limit competition, likening it to major companies seeking to regulate the entire tech sector.

Several lawmakers also noted Congress’s limited expertise in the field. Rep. Jimmy Patronis called for a "kill switch" but warned that legislators lack the technical knowledge to define appropriate safeguards, while Rep. Ralph Norman described the issue as "too complex" for rapid legislative action.

President Donald Trump dismissed the calls for regulation, labeling fears of AI takeover as a hoax and emphasizing that strong leadership, not government rules, is the needed safeguard.

In contrast, Rep. Mike Lawler and Democratic Rep. Josh Gottheimer introduced legislation directing the National Institute of Standards and Technology to develop standards for detecting and controlling rogue AI agents. Lawler emphasized that the goal is to keep people in control without stifling innovation, and Gottheimer outlined provisions for real-time monitoring and rapid disconnection of suspicious AI activity.