OpenAI CEO Sam Altman spoke to a crowd of about 60 lawmakers at a dinner on Monday about the advanced artificial intelligence technology his company produces and the challenges of regulating it.
The wide-ranging discussion, which lasted about two hours, came ahead of Altman’s first testimony to Congress Tuesday at a Senate Judiciary Committee hearing on privacy and technology. IBM Chief Privacy and Trust Officer Christina Montgomery and NYU Professor Emeritus Gary Marcus will also testify at the hearing, which will focus on artificial intelligence supervision.
The dinner discussion comes at a time when artificial intelligence is at its peak, and it has already captured the hearts and minds of Congress. On Tuesday, in the session where Altman will testify, the Senate Homeland Security and Governmental Affairs Committee will hold a separate hearing on artificial intelligence in government. On Wednesday, the House Courts, Intellectual Property and Internet Justice Subcommittee will hold another hearing focused on artificial intelligence and copyright law.
About a half-dozen members who spoke to CNBC outside the Capitol Hill dinner described extensive and informative discussions with Altman, discussing the many fears and hopes for the opportunities AI presents.
Ultraman received high praise from several members.
“I think it’s fantastic,” said Rep. Ted Lieu, D-Calif., the vice chairman of the House Democratic Caucus, who co-chaired with Republican Conference Vice Chairman Mike Johnson (R-La) dinner. “It’s not easy to keep a member of Congress preoccupied for almost two hours. So Sam Altman provided a lot of information, a lot of information.”
“He gave a compelling presentation in real time,” Johnson said. “I think it surprised a lot of members. There was just a standing crowd.”
One of the demonstrations, Johnson said, was to have OpenAI’s generative AI chatbot, ChatGPT, write a bill dedicating a post office to Lieu. He then had Johnson write a speech to be delivered when Johnson introduced the bill in the House.
“It was a beautiful speech,” Lieu quipped.
“It freaked us out too,” Johnson said.
Rep. Haley Stevens, D-Michigan, said that despite her third term in Congress, she had “never been in a meeting like this” and praised Lewis and Johnson for bringing “all of us together.” All representatives of Congress come together to “engage in a topic that is changing our world. “
Rep. Anna Eshoo, D-Calif., co-chair of the Congressional AI Caucus, called Altman “very helpful” and “a pleasure to have thoughtful conversations with.”
“There’s no question where he’s going to back down,” she said, adding that lawmakers have very thoughtful questions to ask.
Eshoo said she had invited Altman to speak at the caucus, but Speaker Kevin McCarthy (R-Calif.) and Minority Leader Hakeem Jeffries (D-N.Y.) insisted on keeping the entire chamber open . Eshoo said she welcomed the opportunity.
“There are a few things you have to understand before you accept or say no,” Eshoo said. “But, it’s like putting a sock on an octopus because it covers everything.”
One of those tentacles has to do with copyright law, something that House Judiciary Committee Chairman Darrell Issa (R-Calif.) has been thinking a lot about.
Issa said he was “very interested in providing more guidance to the Copyright Office,” adding that even if content entirely generated by AI is not copyrightable, guidance is needed on when AI can be copyrightable.
As for Altman, Issa said, in general, “he made it clear that without some legislative and regulatory action, this is not going to move forward, and at the same time, the closing momentum is going to be bad. So, how do you develop guardrails without rubbing side or remove it from the road?”
Rep. Jay Obernolte, R-Calif., who has a graduate degree in artificial intelligence and is a member of the Congressional Artificial Intelligence Caucus, said he discussed with Altman the possibility of regulating precursors to the technology, like the raw materials needed to make nuclear weapons. . Obernolte suggests this could take the form of an international registry that tracks which entities have enough computing power to create advanced artificial intelligence.
Rep. Ro Khanna, D-Calif., whose district spans part of Silicon Valley, said Altman made two important points to members present.
“One is that AI is a tool, not a creature,” he said. “It’s going to help people, not replace people. Second, it’s going to do tasks, not jobs. It’s going to help people do the jobs they have, not replace them. So I think it’s been a cool conversation, helping members understand that. what the tool actually does and helps disprove some of the hype.”
still. There are still some unanswered questions about AI’s vast capabilities, areas where Congress should step in, and OpenAI’s approach to exploiting the technology.For example, some experts have criticized the company choose not to be so candid As for the reason for making its latest large-scale language model, GPT-4, its executives defended it as an important competitive and safety move.
Khanna said the openness of the model was something he had discussed with Altman before, though not at Monday’s dinner.
“The challenge and the value we have to think about is the value of open sourcing it so that other non-incumbents can participate,” Khanna said. “But the danger with open source is that they can fall into the wrong hands. There are trade-offs.”
WATCH: Could China’s ChatGPT clone give it an edge over the US in the AI arms race?