By Richard Cowan
WASHINGTON, Sept 17 (Reuters) - The U.S. Congress should pass legislation by the year-end setting AI safety standards, as an immediate first step toward broad guardrails for an industry whose rapid development has alarmed tech CEOs and the public alike, U.S. Senator Mark Warner told Reuters in an interview.
Warner, vice chairman of the Senate Intelligence Committee and a leading congressional voice on technology and national security, dismissed much of the doomsday talk surrounding
AI but said the time is ripe politically for legislative action, with worries over AI "now at tipping point for Americans."
"I'm not a doomer," the Virginia Democrat said in the interview on Wednesday. "This talk about human extinction is over the top."
He warned, however, of possible trouble ahead, raising "the idea that these swarms of AI agents might jump the fence and take over a banking system or take over a water system or, God forbid, take over a hospital."
A sudden groundswell has risen in favor of federal intervention after OpenAI said in July that an autonomous AI agent went rogue during a security test and hacked into another company, subverting human control.
Warner, who invested in startup companies in wireless telecommunications and other tech sectors before entering politics, said there is an opening for acting before a new Congress is sworn in next January.
"In the short term, I think the only thing we can actually ... get done is a set of safety protocols," he said, adding that UK protocols could be a template for the United States, as could programs that other nations, including China, are pursuing.
TRUMP, REPUBLICANS WARY OVER AI REGULATION
Even that short-term goal, though, is likely to be a heavy lift.
Many Republicans, following the lead of President Donald Trump, are hesitant to thrust the federal government into the AI industry, which is powering U.S. investment and stock market gains.
"I think we need to be careful that we don't get too much government drama," Republican Representative Roger Williams of Texas said in a sidewalk interview near the Capitol on Tuesday. "I think you start out by letting the private sector see how they can compete and make it grow," Williams added.
Republican Representative Ralph Norman of South Carolina, a member of the hard-right House Freedom Caucus, said that while voters want guardrails, "that's going to take a lot of time to work through that." He added: "I'm for limited government."
A bipartisan bill, whose backers range from Democratic Senator Amy Klobuchar of Minnesota to Republican Senator Ted Cruz of Texas, is being refined and Warner said he is reviewing it.
But he stopped short of backing that bill as a first step.
"I think it's well-intentioned, but a lot of it's focused on after-action report, or incident disclosure after action," he said, adding: "We can't wait to have an incident and then look back and see how do we fix it."
TRUMP-XI MEETING PUTS FOCUS ON GLOBAL COLLABORATION
While Trump has taken a hands-off stance toward AI regulation, Washington will be watching closely whether he maintains that position when he meets Chinese President Xi Jinping at the White House on September 24.
U.S. opponents of domestic regulation of AI say any government restrictions would simply give China the upper hand in the AI revolution.
But for many Democrats, including Warner, the U.S. must work together with China, France, the UAE and other nations that are developing AI industries, as the world tries to thwart bad actors.
"This is really one where we have to go back to a level of international collaboration," Warner said, at a time when Trump has been pulling away from such cooperation.
Doing so would benefit U.S. industry, he said.
"It would be in their best interest to have an outside, independent, international veteran in this room."
FEDERAL ROLE FOR AI SAFETY
Warner said eventual legislation could include the creation of a new federal agency overseeing AI.
"Nobody would climb on an airplane or drive a car without some level of outside safety reviews. And the industry would need something like potentially even an antitrust exemption to talk to each other."
With the 100 senators pushing for varying provisions, it could make for a prolonged battle — one that Warner said the country cannot risk.
"I don't think whether we call it a pause or pacing means that we're going to stop innovation" in the AI industry, he said.
One relatively easy provision to legislate, Warner said, would be to impose safety standards as a prerequisite for corporate tax breaks, such as the bonus depreciation.
For the economy, he said there were downsides to AI advancements, such as rising unemployment in some sectors that he said the industry has not addressed while the economy transitions.
But "there will be enormous growth" around blue-collar jobs, such as electricians and welders and healthcare advances. "Everyday Americans and not just a bunch of tech guys out in San Francisco" will benefit, he said.
As he campaigns for a fourth Senate term in this November's midterm election, Warner, 71, warned that Congress often waits for a disaster before addressing problems.
Asked to predict the likelihood of AI safety legislation happening in 2026, Warner said he is not good at making predictions, but made one exception.
"If we don't do something and something happens, there's a 95%-plus probability that Americans are going to be even more pissed off at Congress than at President Trump."
(Reporting by Richard Cowan; Editing by Michael Learmonth and Edmund Klamann)
