President Trump is doubling down on AI development without safety guardrails, framing concerns about artificial intelligence risks as a "hoax" and insisting that only strong presidential authority, not regulation, is needed to manage the technology. Yet this approach is drawing criticism from both industry insiders and policy experts who argue the administration is optimizing for the wrong competition and ignoring genuine threats to American workers and economic stability.

Why Is Trump Rejecting AI Safety Warnings?

Trump has made his position unmistakably clear. In a series of social media posts in mid-September, he dismissed fears about artificial intelligence as alarmism, comparing concerns about AI dangers to what he calls his other favorite targets: climate change and election investigations. He framed AI as "the greatest economic development engine in history" and argued that the only "control or guardrails" the technology needs is a "strong and smart" president.

The president's reasoning centers on three core beliefs. First, he views AI as central to his economic legacy; tech investment accounted for more than one-third of all U.S. economic expansion in the second quarter of 2026, according to forecasting firm ING. Second, he sees AI development as a zero-sum race with China, where slowing American progress would hand victory to a geopolitical rival. Third, he believes that regulation would stifle innovation and hand advantage to competitors abroad.

"We're leading China in AI, and frankly, I want to keep it that way, because whoever wins AI wins," Trump stated in September.

Donald Trump, President of the United States

This stance has elevated AI to the forefront of American politics, even as it creates rifts within Trump's own party. Former Trump adviser Steve Bannon and Senator Bernie Sanders both spoke at a "pro-human" conference on AI dangers in mid-September, with Bannon warning that tech companies should not receive "a carte blanche" and Sanders criticizing Congress for being "asleep at the wheel" on the issue.

Is the U.S.-China AI Race Actually Winnable?

The assumption underlying Trump's strategy is that a decisive, permanent advantage in AI will accrue to whichever nation reaches artificial general intelligence, or AGI, first. But this premise may be fundamentally flawed, according to analysis from the Quincy Institute for Responsible Statecraft.

The evidence suggests that U.S. capability leads over China are shrinking rapidly and cannot be sustained through export controls alone. The performance gap between U.S. and Chinese frontier AI models has narrowed from an estimated 12 to 14 months in 2024 to roughly 2 to 3 months as of 2026, even as Washington steadily increased chip export restrictions. On OpenRouter, the largest model-routing platform, Chinese open-weight models grew from less than 2 percent of token traffic in late 2024 to a majority among top models by 2026.

Most strikingly, recent Chinese releases like DeepSeek's V4 Flash and Moonshot AI's Kimi K3 models are now competitive with leading U.S. systems from OpenAI and Anthropic, while costing roughly one-tenth the price. Even after OpenAI cut prices by 80 percent, DeepSeek's API costs remain 60 percent lower per task.

Unlike nuclear weapons, which can be physically secured and remain strategically dominant for decades, AI models are becoming interchangeable commodities that cross borders in minutes. As one analysis noted, there is no clear "finish line" in AI development that would grant permanent dominance to either nation.

What Are the Real AI Security Threats Trump's Strategy Overlooks?

By framing AI safety as a distraction from competition with China, the Trump administration may be missing more immediate and dangerous risks. Policy experts argue that the most dangerous AI models are not necessarily the largest frontier systems built by nation-states, but smaller, specialized models that could be created or weaponized by non-state actors for deploying biological, chemical, or cyber weapons.

The administration's focus on outpacing China also prevents the international cooperation needed to establish shared safety standards and harm-prevention measures. Treasury Secretary Scott Bessent has stated that "if China were to pull ahead of us on AI, then nothing else matters," a framing that subordinates all other concerns, including worker protection and financial stability, to competitive advantage.

How Should Policymakers Balance AI Innovation and Safety?

• Reject National Security Exemptions: The U.S. should not grant AI labs exemptions from ordinary market discipline and legal regulation simply because AI is framed as a national security issue. Doing so removes accountability mechanisms that protect workers and the broader economy.
• Establish Cooperative Safety Standards: Rather than viewing AI development as a winner-takes-all race, the U.S. and China should pursue cooperative measures around shared harm and safety standards, particularly for smaller specialized models that pose asymmetric risks.
• Prioritize Worker Protection: Policy should focus on safeguarding American workers from displacement by automation, rather than assuming that maximum AI acceleration will automatically benefit the broader population.

The political calculus around AI regulation remains uncertain. With midterm congressional elections looming, both Republicans and Democrats are wary of taking positions that could alienate voters who are increasingly skeptical of AI's promised benefits. Republican Speaker of the House Mike Johnson has already backed away from earlier openness to congressional action, instead echoing Trump's "AI hoax" rhetoric.

However, experts note that public opinion could shift rapidly if a high-profile AI-related disaster occurs. As tech firm Appian's head Matt Calkins observed, "It is possible that events in the future might electrify the public, and then positions and talk might not be so cheap". The moment when AI warnings turn into a visible crisis could reshape the political landscape and force a reckoning with the safety concerns Trump is currently dismissing.