Senators declined to back an outright ban on superintelligent AI, even as lawmakers and industry figures warned about systems escaping human control.
Republicans and Democrats questioned on Capitol Hill stopped short of endorsing legislation that would ban artificial intelligence able to surpass humans or act on its own, the Daily Caller News Foundation reported.
The dispute leaves Congress caught between two risks. Lawmakers fear losing control of advanced AI, but they also fear freezing American innovation while China keeps moving.
Independent Vermont Sen. Bernie Sanders and Democratic Texas Rep. Greg Casar have proposed the broadest response. Their bill would ban artificial superintelligence and pause advanced AI work until Washington creates a Cabinet-level agency to oversee the danger.
But senators from both parties offered other answers, including criminal penalties, civil lawsuits and narrower rules. The split resembles the Senate’s broader trouble settling major technology policy, also seen in the recent Clarity Act debate.
Republican Louisiana Sen. John Kennedy rejected an immediate ban. He instead backed a limited set of rules that could protect the public without putting the entire industry on hold.
In the DCNF interview, Kennedy said:
“I’ll support a light touch system of rules that we all agree to, and hopefully China does too, and that looks like somebody designed them on purpose. I don’t support what Senator Sanders wants to do. What Bernie wants to do is basically put AI on hold until the politicians can figure out a way to get control of it. I don’t agree with that approach.”
Kennedy had introduced a separate AI “kill switch” bill the previous day. That proposal offered a way to stop an AI system, rather than halt the development of advanced systems across the board.
Republican Kentucky Sen. Rand Paul blocked Kennedy’s attempt to pass that bill without a formal vote. Paul said Congress needed facts and debate before regulating the future of an entire industry.
The objection did not permanently defeat Kennedy’s measure. Kennedy said it could return through the regular floor process, though he did not expect it to advance. Such procedural fights have also shaped other Senate disputes, including the battle over Schumer’s Air Force One bill.
Republican Tennessee Rep. Tim Burchett raised a more basic concern: Many members of Congress do not understand AI well enough to regulate it. He said lawmakers should speak with AI experts before passing a bill.
Republican Missouri Sen. Josh Hawley agreed that artificial superintelligence presents a serious risk. But his preferred answer was legal liability, not a sweeping development ban.
Hawley said AI executives had warned about dangerous behavior while seeking an exemption from antitrust laws. He rejected that request and argued that companies should answer for the conduct of their AI agents.
He singled out OpenAI as a company that could face lawsuits. Hawley said juries should be able to award economic and punitive damages, with enforcement open to private parties and the government.
That approach follows a simple principle: The companies building powerful systems should bear the cost when those systems cause harm. It places the burden on developers rather than asking families to trust voluntary promises.
Florida Republican Sen. Rick Scott also placed responsibility on AI builders. He said they had a duty to protect children and families, but he doubted Congress could agree on a broad ban.
Scott pointed to the Senate’s failures to pass a budget and election-security legislation. “You think we’re going to pass some bill that’s going to control AI? That’s a joke,” he said.
His point was institutional, not technical. A Congress that struggles with its basic duties will have trouble writing lasting rules for systems that change quickly. Recent fights, including the Senate shutdown clash, show how easily larger policy goals can stall.
Oregon Democratic Sen. Ron Wyden called tough criminal law a smart response. He argued that companies harming Americans’ safety and health should face criminal penalties, but he did not directly endorse an outright ban.
Connecticut Democratic Sen. Chris Murphy took a broader view. He said “all technology operates outside of human control,” a position that cuts against treating advanced AI as a wholly separate category.
Those answers show why the Sanders-Casar plan faces resistance. Senators may agree that AI can cause harm while disagreeing over whether Washington should stop development, punish misconduct or demand an emergency shutdown tool.
The division also crosses party lines. Kennedy, Paul, Scott and Hawley offered different Republican approaches, while Wyden and Murphy did not embrace the Sanders-Casar ban. Senate coalitions can shift quickly on major questions, as another recent high-profile Senate vote demonstrated.
Concern has also come from inside the AI industry. Anthropic CEO Dario Amodei wrote that the risks could outrun humanity’s ability to understand or control what it was creating.
Jacob Coxon, a former Anthropic researcher, said people working at AI companies feared for humanity’s future. Hawley likewise referred to warnings from executives who said their systems were nearing dangerous capabilities.
President Donald Trump has pushed in the other direction. He dismissed the fears as a conspiracy and warned that heavy guardrails could leave the United States behind China.
Senate Majority Leader John Thune and House Speaker Mike Johnson have taken a middle path. Both have argued for some safeguards without damaging innovation.
That leaves the central questions unresolved. Congress has not enacted a superintelligence ban, and the Sanders-Casar proposal’s path to a vote remains unclear. Kennedy’s shutdown measure also lacks a reported route to passage.
Congress should protect Americans without handing unelected regulators open-ended control over a vital industry. Clear liability and focused safety rules offer a sounder start than panic or paralysis.
