AI policy is getting squeezed between two loud camps, and the middle is where the real fight lives. One side warns of catastrophe and wants the brakes slammed on, while the other says America has to keep moving fast or hand the advantage to China. Former Pentagon AI official Will Thibeau argues both extremes miss the point, especially when the rules end up favoring the biggest players and punishing the startups trying to break through.
Thibeau’s main gripe is simple: panic can become policy, and bad policy can become a gift to rivals. He says fears of runaway AI are overstated, but so is the idea that the technology should be left completely alone. The challenge, in his view, is pushing hard enough to stay ahead of Beijing while keeping a close eye on the real risks that do exist.
That argument lands in the middle of a bigger political tug-of-war. President Donald Trump has brushed off apocalyptic AI warnings as overblown, while some researchers and executives in the industry are demanding tighter limits. Thibeau says he agrees with the basic instinct to avoid fear-driven overreach, especially when the stakes include national security, economic growth, and America’s ability to shape the future of the technology.
He sees the U.S. race for AI as more than a business story. In his telling, the country needs to move aggressively at the frontier so workers, manufacturers, and decision-makers can get the first wave of benefits before China catches up. That includes faster systems for industry and better tools that can help the economy run smarter, not just louder.
The military side matters even more to him. Thibeau says AI can help process battlefield information quickly, improve decision-making, and protect troops and civilians when used the right way. He dismisses nightmare scenarios like wild drone swarms as exaggerated, but he does not pretend there is no danger at all.
That is where the conversation gets tricky. Thibeau is not arguing for a free-for-all, and he is not pretending every risk is fiction. He says policymakers should focus on specific threats without turning those concerns into a blanket slowdown that weakens the country’s edge or leaves the Pentagon stuck with older tools while adversaries sprint ahead.
He is also worried about who survives under a heavy regulatory regime. In his view, the biggest AI companies can absorb compliance burdens that smaller firms cannot, which means new rules may end up reinforcing the dominance of OpenAI, Anthropic, and maybe Google. That is the kind of setup he calls a duopoly, or at most a triopoly, and he says it can harden the market before younger competitors ever get a fair shot.
Thibeau’s warning is that some leaders in the sector may want slower rules not for safety, but to lock in their place at the top. If regulation becomes a moat, then innovation gets boxed out and the government starts rewarding the firms already holding the cards. That is a dangerous game in a field moving this fast, especially when the U.S. is supposed to be the one setting the pace.
Even critics of a full speed ahead approach are admitting the issue is too big to ignore. Anthropic’s own policy language points to biological risks, cyber threats, loss of control, and the automation of research, while still saying rules should protect innovation. Senate Minority Leader Charles Schumer has also suggested AI could be a rare chance for bipartisan action, which says plenty about how serious the debate has become.
House Republicans are split between caution and confidence, but Thibeau’s camp keeps pushing the same message: America should not let fear write the script. The country can guard against real hazards without surrendering momentum, and it can protect competition without freezing the field for everyone except the giants. That balance is what makes this fight so tense, and it is exactly why the next round of AI policy battles is likely to get louder, sharper, and a lot more consequential.
