Mr. ChatGPT goes to Washington: OpenAI CEO Sam Altman will testify before Congress on the dangers of AI

(CNN) OpenAI CEO Sam Altman urged lawmakers to regulate AI during a Senate hearing on Tuesday, describing the technology’s current boom as a potential “printing moment” but one that requires safeguards.

“We believe that regulatory intervention by governments will be critical to mitigate the risks of increasingly powerful models,” Altman said in his opening remarks to a Senate Judiciary Subcommittee.

Altman’s emergence follows the wild success of his company’s chatbot tool ChatGPT, which has renewed the arms race over artificial intelligence and raised concerns among some lawmakers about the risks posed by the technology.

Sen. Richard Blumenthal kicked off the hearing Tuesday with a fake recording of his voice, illustrating the potential dangers of the technology. The recording, which included notes written by ChatGPT and an audio of Blumenthal’s voice produced using recordings of his actual speeches, argued that AI could not be allowed to emerge in an unregulated environment.

Blumenthal explained that while ChatGPT provided an accurate reflection of the views of a real legislator, it could easily have produced “an endorsement of the surrender of Ukraine or the leadership of Vladimir Putin.” This he said, “It would have been really scary.”

A growing list of tech companies have deployed new AI tools in recent months, with the potential to change the way we work, shop, and interact with each other. But these same tools have also drawn criticism from some of the biggest names in tech for the potential to disrupt millions of jobs, spread misinformation and perpetuate bias.

In his remarks Tuesday, Altman said the potential for AI to be used to manipulate voters and target disinformation is among the “areas of my greatest concern,” especially because we’re going into next year’s election and these models are improving. “

Altman said Tuesday that one way the US government can regulate the industry is to create a licensing regime for companies that work on the most powerful AI systems. Altman said that this “combination of licensing and testing requirements” can be applied to “the development and release of AI models above the capability threshold.”

Also testifying Tuesday will be Christina Montgomery, IBM’s vice president and chief privacy and trust officer, as well as Gary Marcus, a former New York University professor who has described himself as a critic of AI “noise.”

Montgomery warned of creating a new era of “move fast and break things,” the old mantra of Silicon Valley giants like Facebook. “The age of AI can’t be another era of ‘going fast and breaking things,’” Montgomery told lawmakers. However, she said, “we don’t have to hold back innovation either.”

Altman and Montgomery also said that AI may eliminate some jobs, but it creates new ones.

“There will be an impact on jobs,” Altmann told Blumenthal. “We’re trying to be very clear about that, and I think it’s going to require a partnership between industry and government, but often taken by government, to figure out how we want to mitigate that. But I’m very optimistic about how great the jobs future will be.”

As CEO of OpenAI, Altman has become, perhaps more than any other figure, the face of a new set of AI products that can generate images and text in response to user prompts.

Altman’s comments come a day after he met more than 60 House of Representatives for dinner. The bipartisan caucus, which has an even split of Republicans and Democrats, saw Altman demonstrate various uses of ChatGPT “much to his amusement,” according to a person in the room who described lawmakers as “energized” by the event.

The person added that most attendees widely acknowledged that regulation of AI would be necessary.

California Democratic Rep. Ro Khanna, whose district includes Silicon Valley, said Altman stressed over the dinner that AI is a tool, not a “creature” and that AI “can help with tasks, not jobs.”

“Altman’s most useful contribution was cutting down on the noise,” Khanna told CNN.

In a reflection of how artificial intelligence has invaded Congress, even as the Judiciary Subcommittee was questioning OpenAI and IBM, the Senate Homeland Security and Governmental Affairs Committee was holding a separate, simultaneous hearing on the use of AI in government.

Earlier this month, Altman was one of several tech CEOs who met with Vice President Kamala Harris and, briefly, President Joe Biden as part of a White House effort to stress the importance of ethical and responsible AI development.

In interviews this year, Altman has presented himself as someone who understands the dangers posed by artificial intelligence and is even “a little afraid” of the technology. He and his company have vowed to move forward responsibly.

Others want Altman and OpenAI to move more carefully. Elon Musk, who helped found OpenAI before dropping out of the group, has joined dozens of technology leaders, professors and researchers in signing a letter calling for AI labs like OpenAI to stop training the most powerful AI systems for at least six months, citing “profound risks to society.” and humanity.”

Altman said he agreed with parts of the letter. “I think treading with caution and the increased rigor of safety issues is really important,” Altman said at last month’s event. “The message I don’t think was the best way to counter it.”

— CNN’s Jennifer Korn contributed to this report.