Leaders Debate the Dangers and Future of Artificial Intelligence

Top technology leaders and government officials gathered in Washington, D.C., on Wednesday for POLITICO's Decoded Summit to talk about artificial intelligence, or AI. The event brought together tech company executives, members of Congress, and former White House officials to debate how powerful AI is becoming — and whether the U.S. needs new rules to keep it safe. The discussions came just days after the heads of two major AI companies called for slowing down the development of the technology, warning it could be very dangerous. The summit highlighted big disagreements about how serious the risks really are and what steps leaders should take next.

Some speakers at the summit strongly pushed back on warnings that AI could be a threat to humanity. Former White House AI adviser David Sacks called those warnings a 'hoax,' saying the media was scaring the public unnecessarily. He praised President Donald Trump for speaking out against calls to slow AI development. Sacks said that Trump's willingness to challenge the narrative 'creates the space to then dissect the narrative and find out all the ways in which it is flawed.'

House Energy and Commerce Committee Chair Brett Guthrie compared the AI fear to past 'doomsday scenarios' about climate change, saying many of those predictions did not come true. LinkedIn co-founder Reid Hoffman took a softer position, calling the threat of AI destroying humanity 'remote and controllable.' But even Hoffman admitted there are real concerns, especially around cybersecurity and the potential misuse of AI to create dangerous biological weapons. 'Cyber is very real, and it's here now,' Hoffman said.

Not everyone at the summit was so calm about the risks. Sarah Heck, who works on public policy for Anthropic — one of the companies that called for a slowdown — argued that the fears are serious and deserve action. 'We do think it's existential to do this. We do think we have to balance the risk now,' she said. She pointed out that AI systems are becoming more powerful at a faster and faster pace, making it urgent to act now rather than later.

A big part of the debate focused on whether putting strict safety rules on AI would help China get ahead of the United States. Many Trump administration officials and Republican lawmakers argue that slowing down AI development would give China an advantage in a global technology race. 'We absolutely can't surrender,' said Guthrie at the summit. He believes stopping AI development would be like giving up in a competition the U.S. needs to win.

But not everyone agreed. Representative Greg Casar of Texas argued that the competition with China is not a reason to ignore safety. 'The AI race is not an arms race that you win,' he said. 'It's an arms race that we need to end, and we need to protect the American people.' Casar is working with Senator Bernie Sanders on a bill that would ban the creation of artificial superintelligence — a type of AI that would be smarter than any human.

Anthropic's Heck offered a middle-ground view, saying the U.S. needs to stay ahead of China specifically so it can set the safety rules for the rest of the world. She also argued that the U.S. should limit the sale of AI chips to China to keep its technological lead. Her point was that the country in front gets to decide what safety guardrails look like — and the U.S. should make sure that country is itself.

Speakers offered a wide range of ideas for making AI safer, even those who doubted the worst-case scenarios. Casar called for creating a new government agency focused entirely on AI safety, comparing it to the Nuclear Regulatory Commission that oversees nuclear energy in the U.S. He said that if top AI company leaders are comparing AI to nuclear weapons, then the government should treat it just as seriously. Former Indiana Governor Eric Holcomb pressed for programs to help workers whose jobs might be replaced by AI, including job retraining and support for people whose careers are disrupted.

Sacks took a different approach, saying the laws that already exist are enough to handle AI dangers. He agreed with a statement by former Federal Trade Commission Chair Lina Khan that AI companies should be held responsible — through product liability laws — when their systems cause harm. Meanwhile, Clem Delangue, the CEO of AI company Hugging Face, said the government must make sure AI companies are responsible for damage caused by their AI systems. His company was the target of the first known fully autonomous cyberattack — meaning an attack carried out entirely by an AI without a human giving it instructions.

The summit also touched on politics beyond AI. When asked about former Vice President Kamala Harris possibly running for president in 2028, Hoffman said it would be 'a very challenging path for her.' He also said he might be open to supporting Representative Alexandria Ocasio-Cortez, calling her 'much more pragmatic than people give her credit for.' Hoffman named several Democratic governors — including Gavin Newsom of California, Josh Shapiro of Pennsylvania, Andy Beshear of Kentucky, and Wes Moore of Maryland — as leaders having good conversations about AI safety. The summit made clear that AI is not just a technology issue — it has become a major political topic that will shape elections and government policy for years to come.