Trump says AI can self-regulate. Even some top Republicans don’t trust that
Want to bookmark your favourite articles and stories to read or reference later? Start your Independent Membership today.
For the past month, as Americans’ anxieties about artificial intelligence have increased, President Donald Trump has repeatedly said there should be no legislative guardrails, and called the panic around them a “hoax.” Rather, he’s argued that he alone can rein them in and argued that if the United States doesn’t go all-in on AI, China will win the AI race.
This week, he even attempted to re-brand AI as “super intelligence.” And he brought the nation’s leading tech giants to the White House and had them sign a “morally binding” and “self-policing” accord. Trump hosted AI leaders such as Anthropic’s Dario Amodei, Google’s Sundar Pichai, Meta’s Mark Zuckerberg and SpaceX’s Elon Musk, saying the companies signed a “morally binding” accord to “self-police” AI.
But amid all of these declarations, even the most MAGA of Republicans are saying that the voluntary regulations are not enough and that there needs to be oversight of AI, SI, or whatever they want to call it, beyond Trump and the tech titans themselves.
Sen. Josh Hawley (R-Mo.), for example, convened a Senate Homeland Security Subcommittee on Disaster Management on rogue AI and cyber threats. Hawley, perhaps the foremost critic of big tech on the Republican side, has had to walk a fine line between his criticism of AI and the fact that Trump has said the only guardrail AI needs is a “STRONG AND SMART (High IQ!) PRESIDENT.”
“He brought the leaders of these companies together for the first time to sign some voluntary commitments, which is a heck of a lot more than Congress has done,” Hawley told The Independent.
“So I think that it's good to have, No. 1, the companies acknowledge that they have some responsibility here,” Hawley said. “This is really the first time that they've done that, and I think Congress needs to do something.” But the Missouri Republican also said incentives needed to be put in place to force AI companies to pay if they develop models that misbehave.
This comes after OpenAI had to apologize after an autonomous AI agent the Services Australia Medicare Statistics Reporting Service, a public health care data portal in Australia.
Sen. Joni Ernst (R-Iowa), an ally of the president who is retiring, also said the self-regulation could only be temporary.
“For now, they can choose to self-regulate, but it’s all voluntary,” she told The Independent after the subcommittee hearing.
So far, no major legislation has passed on reining in AI yet, but there are some trial balloon pieces of legislation in the House of Representatives. Sen. Mark Kelly (D-Ariz.) has his own piece of legislation called the Make AI Work for Americans Act, which would set up a federal trust fund to support workers amid major changes AI could make to the global economy.
“In my prior career, when you're flying something like the space shuttle, there's like catastrophic things that could happen,” Kelly, a former pilot and astronaut, told The Independent. “They might be low probability, but you still try to deal with it.”
He also cast doubt on Trump saying earlier this month in an interview with GB News that humans will have a “little gear” to prevent AI from running rampant.
“I don't even know what that is,” Kelly said. “I don't think he knows what that is. We have to put guardrails in place, and it's always better to do that sooner rather than later.”
Earlier this month, former Anthropic researcher Jacob Coxon set off alarms when he announced his resignation and said that people working on AI genuinely fear it “could kill us all by the end of the decade.” In response, Sen. Bernie Sanders (I-Vt.) and former Trump adviser Steve Bannon participated in a summit focused on human-centered AI.
“Yes, I have absolute confidence that the wealthiest, most greedy people in the world, who are out to make trillions of dollars, will regulate themselves,” Sanders told The Independent. “I don't think so.”
But Trump has proven allergic to any safety measures around rogue AI agents, instead focusing on rebranding it “Super Intelligence.”
Sen. Tommy Tuberville (R-Ala.), who is leaving the Senate to run for governor, said AI companies will will have to do some self-regulation.
“At the end of the day, we're going to there's going to be something done up here,” he told The Independent. “But everybody's got to take care of their own business. I mean, they'll self-destruct if they don't do their self-regulation to a certain point.”
But Sen. Jon Ossoff (D-Ga.) the youngest senator who is running for re-election, said the president has a vested interest in opposing guardrails on AI, pointing out the president has a financial stake in many companies working on AI.
“The first family is so heavily invested in the industry, the conflicts are so immense, and the president's judgment is so poor that the American people are rightly concerned that Donald Trump has neither the capacity nor the incentive to protect them from the risks arising from this technology,” he told The Independent.
But despite the concerns, there is little likelihood the Senate will take up any action of AI by the end of this year. The Senate left late on Wednesday evening and will not return after the midterm election. After which time, it will be in for a limited amount of weeks because of the holiday season before the new Congress convenes at the beginning of the year.
