More than 20 tech and civil society leaders, including the chief executives of five of the 10 biggest U.S. companies, appeared at a closed-door Senate meeting on Wednesday to shape how artificial intelligence is regulated.

The meeting, which was organized by Senate Majority Leader Chuck Schumer, included a prestigious, and possibly combustible, mix of personalities with diverging views on how to write the rules for AI. The CEOs of Alphabet Inc., Microsoft Corp., Meta Platforms Inc. and OpenAI were invited to appear alongside rivals and industry critics to discuss possible guardrails for AI that balance the risks and rewards of the technology.

Areas of disagreement were apparent throughout the morning session, according to several people who were in the room. Meta CEO Mark Zuckerberg, OpenAI CEO Sam Altman and Microsoft co-founder Bill Gates offered diverging views on the risks of open-source AI research, according to people in the room. Tesla Inc. CEO Elon Musk was also met with pushback from Berkeley researcher Deb Raji for appearing to downplay concerns about AI-powered self-driving cars, according to one of the people in the room.

Apart from policy disputes, there is also personal friction among some of the attendees. Musk (the world’s richest man) previously proposed a cage fight with Zuckerberg (the world’s 10th richest man) and took to his own social media platform to criticize Gates (the world’s 4th richest man) for allegedly shorting Tesla stock.

“That’s probably the worst wedding to try to do seating for,” said Humane Intelligence CEO Rumman Chowdhury, one of the panelists. She said there was zero interaction between Musk and Zuckerberg, who were seated on opposite ends of a long table that ran the width of the room. Despite some participants holding different views, Chowdhury described the discussion as productive overall.

The high-profile summit comes as part of Schumer’s broader efforts to get the Senate up to speed on the rapidly evolving technology. While AI could revolutionize industries from health care to logistics, it also poses a number of risks, including spreading misinformation, amplifying bias and displacing jobs. Some industry critics have previously raised concerns about the potential for tech executives to have too much influence over legislation.

Schumer personally called many of the CEOs and other panelists to ask them to come, according to two people who work with different meeting participants. Amazon was invited but declined to come, citing a board meeting, according to two people familiar with the matter.

The forum started just after 10 a.m. ET. On his way into the summit, Altman, who has emerged as a central figure in discussions about regulating the AI sector, said: “This is sort of an important and urgent and in some ways unprecedented moment.” Musk, meanwhile, called AI a double-edged sword in off-the-cuff remarks at the meeting, telling US senators that the technology can be a tremendous source for good but warning about risks to civilization, according to a person in the closed-door session.

Schumer opened the meeting Wednesday noting the unusual format and argued that such a “unique” approach is warranted, given the importance and potential power of this new technology.

“Today, we begin an enormous and complex and vital undertaking: building a foundation for bipartisan AI policy that Congress can pass,” Schumer said to open the meeting. “This is going to be one of the hardest tasks we undertake, because AI is so complex, will impact nearly every area of life, and is evolving all the time.”

In corporate blog posts and White House meetings over the past year, some of the companies represented on Wednesday have proposed a new government agency to regulate AI, a licensing regime for the most powerful AI models and transparency requirements for AI-generated content. Other companies like International Business Machines Corp. have pushed back on some of these suggestions and offered alternative frameworks for legislation.

“An AI licensing regime would be a serious blow to open innovation and risks creating a form of regulatory capture,” IBM CEO Arvind Krishna said in a statement. “This would inadvertently increase costs, hinder innovation, disadvantage smaller players and open-source developers and cement the market power of a few players.”

Most of the executives in the room have, in one way or another, urged policymakers to focus on setting rules for the way AI is used rather than for the actual development of the technology. Many of the participants stressed that the U.S. should play a leading role in shaping global governance of AI, and some referenced China’s tech advancements as a specific concern, according to people in the room. At one point, Schumer asked for a show of hands if the participants thought that the US government should plan an oversight role, and nearly every hand in the room went up, according to the people.

In addition to the CEOs, leaders from organized labor and civil society were also in the room Wednesday, warning about the risks that AI poses for workers, mental health and democracy. Liz Shuler, president of the AFL-CIO, and Meredith Stiehm, president of the Writers Guild of America West, were among those invited to attend.

Some senators have already questioned the practice of giving tech executives a platform to explain how they would like to be regulated, especially behind closed doors. Massachusetts Democrat Elizabeth Warren on Monday said the best way to “hear from the tech billionaires is a hearing” in public where people can push back on their arguments.

South Dakota Republican Mike Rounds, who worked with Schumer to organize the set of briefings to educate senators on AI, said the goal of this format is to allow for “lots of different opinions in a very short period of time.” He said the aim is to “be as frank as possible, not necessarily grandstanding in front of a camera.”

“They’ll give us their opinions, we’ll take that with a grain of salt, but most certainly we’ll all come out with more information than we had going in” Rounds told Bloomberg News Tuesday.

Schumer has tried to impart a sense of urgency on moving forward with legislation while also deferring to the committee process, which can take some time. Speaking to reporters during the lunch break, Schumer said Wednesday’s insight forum will give committees the material they need to actually write legislation.

Maya Wiley, the president and CEO of the Leadership Conference on Civil and Human Rights, and one of a handful of civil society figures in the room, said she was encouraged by the unanimity on regulating AI, but crafting concrete policies requires further discussion. Schumer is expected to host at least eight more AI forums, touching on issues from national security to privacy.

“The tension points come in how you balance speed of the technology and figuring out how you incorporate the safety, the transparency, and the accountability in the front end so people aren’t harmed later,” Wiley said.

--With assistance from Steven T. Dennis and Kailey Leinz.