The people building and governing A.I. increasingly fall into camps: Those who want to slow down and regulate the technology’s development, those racing to speed it up and those in between or with shifting stances.
In the last week, calls to slow the development of artificial intelligence have moved to the center of the national agenda. What began as a philosophical question (how dangerous might this technology become?) has hardened into increasingly entrenched rival factions.
One side is warning that the technology is advancing faster than humans can control it. Another is treating the race for dominance against competitors at home and abroad as the only risk that counts. A third group in the middle is pressing for guardrails that would neither halt the development of A.I. nor wave it through, and a handful of prominent voices have shifted sides as the technology has changed.
The divisions are increasingly defining the debate over A.I., and they do not follow familiar lines. The dispute has split Silicon Valley, set technology executives against President Trump, who has rejected their calls to slow down and united populists on the left and right who agree on little else but that A.I. should be reined in.
The outcome will shape more than just the industry. Whichever side prevails will influence how A.I. enters daily life and which nation sets those rules for everyone else. So far, no camp has triumphed and the technology continues to advance.
The New York Times mapped many of the people driving that fight: building the systems, funding them or deciding whether the government steps in.
Executives
Slow down
Dr. Amodei left OpenAI, a leading A.I. lab, to start Anthropic in 2021, which he has positioned as a safety-first firm. He has publicly warned about the potential catastrophic outcomes from advanced A.I., including the creation of bioweapons. He has repeatedly pushed for government regulation. Most recently, Dr. Amodei called for a global slowdown of A.I. development, laying out a plan in a 3,800-word essay to rein in the technology.
Shifted
Mr. Musk has long put the odds of A.I. “going bad” at roughly 10 to 20 percent, calling it more dangerous than nuclear weapons. He has argued for proactive regulation before it’s too late. Yet he founded xAI and spent heavily to elect President Trump, whose administration he briefly joined to lead its government-cutting and dergulation effort. He said in a social media post on Saturday that “Dario is right” about the need to slow down A.I. development.
In the middle
In 2010, Dr. Hassabis helpd found DeepMind, an A.I. lab that was later acquired by Google. He has said A.I.’s risks are serious enough that they cannot be ignored, comparing the stakes to climate change. He has called for “smart regulation” and oversight. Dr. Hassabis, who has described his viewpoint as cautious optimism, publicy supported Dr. Amodei’s essay to rein in A.I. while adding the details still need to be worked out.
Mr. Nadella has led a valuable partnership between Microsoft and OpenAI and, more recently, Microsoft’s in-house A.I. tools. In a social media post over the weekend, he said he backed “deliberate pacing” of A.I. development, pushing for human oversight of advanced systems. He has said that superintelligence is not worth pursuing unless it stays under human control and serves people broadly — while simultaneously pushing wide industry adoption and an open ecosystem.
Mr. Pichai has said for years that A.I. is “too important not to regulate well” and has acknowledged the technology’s risks. But he prefers rules tailored to specific uses over heavy restrictions, and has repeatedly framed the larger danger as failing to capture A.I.’s benefits — positioning Google as a builder that wants guardrails without a slowdown.
Mr. Zuckerberg’s believes that A.I concentrated in the hands of a select few companies is the real danger and broad access is the safeguard. He has warned that rules slowing U.S. model releases would hand the lead to China. Mr. Zuckerberg, however, acknowledges safety concerns, and has signaled Meta may not openly release its most powerful future models.
Mr. Wang joined Meta in 2025 to lead its chief A.I. initiatives and the development of Meta’s personal digital assistant agent, Muse. He is a co-author of a strategy paper framing superintelligence as the most precarious technology since the nuclear bomb and has argued for a national security deterrence approach.
Speed up
Mr. Huang runs the company whose chips power most of the A.I. industry. He has become one of tech’s most prominent proponents of the technology, arguing that no company should be asking the government for more A.I. regulation. He has also clashed publicly with Dr. Amodei over predictions of job losses and export controls of chips.
A longtime venture investor, Mr. Sacks was installed by the Trump administration as its top A.I. and crypto official. He has used the term “Doomer Industrial Complex” to dismiss calls for increased safety built on fearmongering. He has opposed the proliferation of state-level A.I. laws in favor of a light-touch federal approach. He criticized Dr. Amodei’s recent call to “pace” A.I. development, adding that A.I. labs set the technological frontier themselves.
A pioneering venture capitalist, Mr. Andreessen is one of the sharpest voices for speeding up A.I. development. His 2023 “Techno-Optimist Manifesto” explicitly listed “existential risk” among the enemies of progress, and he has likened A.I. regulation to Prohibition. Mr. Horowitz is the co-leader of a16z and its policy pushes, and has taken similar anti-regulation stances.
Politicians
Regulate
Mr. Sanders has become one of the loudest critics on Capitol Hill, casting the technology as a tool of “oligarchs” that threatens working people. Mr. Sanders has proposed a moratorium on new A.I. data-center construction, a one-time 50 percent tax on leading A.I. firms funneled into a public wealth fund, a permanent ban on superintelligence and a pause on advanced development until a federal regulator exists.
In an interview on ABC’s “This Week,” Mr. Jeffries pressed for “decisive action” on A.I. and criticized Mr. Johnson for moving too slowly. In December, he started a House Democratic Commission on A.I. and the Innovation Economy, and has argued Congress should not let executives regulate their own technology, though he has so far declined to commit to specific restrictions.
Mr. Hawley has become one of the Senate’s most prominent Republican A.I. skeptics. He is a co-sponsor of legislation in 2025 that would classify A.I. systems as products, allowing for legal liability claims when an A.I. system causes harm. The chairman of a Senate subcommittee on disaster management, he sent a letter to Mr. Altman last week requesting more details on the Hugging Face attack as part of a formal probe.
The Senate’s No. 2 Democrat and its top Judiciary Democrat, Mr. Durbin pushed for a bill targeting A.I.-generated non-consensual intimate imagery and has led hearings on chatbots’ harms to children. He co-sponsored the legislation with Mr. Hawley that would classify A.I. systems as products, allowing for liability claims when a system causes harm.
As the Senate’s top Republican, Mr. Thune told reporter on Monday he favored a “light touch” on A.I. regulation. He called for guardrails aimed only at the most consequential threats without slowing innovation or ceding ground in the A.I. race. Mr. Thune is helping to craft a bipartisan A.I. bill that would establish some safety and security guardrails for the riskiest models.
Don’t Regulate
Mr. Trump has made A.I. dominance a priority, framing it as a race the U.S. must win against China. He has emphasized rapid buildout — including the electricity needed to power the data centers that run A.I. — and support for open models. On Monday, he rejected calls from A.I. executives to limit the technology. In a call with Mr. Huang of Nvidia that was played live at a conference, Mr. Trump suggested he was unwilling to intervene in A.I. development.
Mr. Johnson has acknowledged A.I.’s risks are rising but has resisted moving quickly. In an interview on Sunday with NBC’s “Meet the Press,” he said the House should not hastily pass regulations that could hand China an edge. He said on Tuesday that Mr. Trump would bring together all the “platform providers” in a meeting in the next week, but dismissed fears of A.I. wiping out humanity.
