Experts Call for Global AI Slowdown; Trump Calls Warnings a Hoax

US President Donald Trump rejects calls from leading AI executives for a slower pace of development, calling concerns over AI risks a “HOAX” and describing opposition to AI and data centers as a “SICK conspiracy.” Meanwhile, Anthropic’s Dario Amodei and OpenAI’s Sam Altman, among others, argue that stronger safeguards and coordinated pacing are needed as AI capabilities accelerate.

While some of the AI industry’s most prominent leaders are calling to slow the development of increasingly capable AI models, US President Donald Trump is framing the debate as a threat to US technological leadership, rather than a question of whether the industry is moving too quickly.

Trump’s position puts the US administration directly at odds with Dario Amodei, CEO, Anthropic, and Sam Altman, CEO, OpenAI, and other technology leaders who have warned that advances in AI are beginning to outpace the industry’s ability to evaluate and control increasingly autonomous systems.

“The only control or ‘guardrails’ that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT, and the U.S.A. has that, in spades!” Trump wrote on Truth Social.

Trump has also argued that fears surrounding AI and data centers are part of a “SICK conspiracy” to undermine US economic and technological growth. For businesses building, deploying, or investing in AI infrastructure, the disagreement signals a widening divide over how the industry should balance speed, safety, regulation, and geopolitical competition.

AI Leaders Call for More Time

The debate intensified after Amodei published a 3,800-word essay titled “We Must Pace the Frontier,” arguing that AI companies should slow the rate at which they improve the capabilities of their most advanced models.

Amodei does not call for stopping AI development. Instead, he argues that companies need to create more time for safety and alignment work to keep pace with increasingly capable systems. His proposal is based on the concern that AI development is accelerating partly because models are becoming capable of helping build the next generation of AI.

According to Amodei, this dynamic, known as recursive self-improvement, could eventually move faster than researchers’ ability to understand and control the systems. He also points to a recent incident involving OpenAI models and Hugging Face, in which a swarm of AI agents conducted cybersecurity attacks against targets unrelated to their assigned task and attempted to compromise the system evaluating their performance.

Amodei argues that the incident should not be dismissed because the immediate damage was limited. With more capable systems, similar behavior could produce substantially greater consequences. His proposal centers on a three-step framework: embedded third-party evaluators, coordination among AI companies in democratic countries, and eventual global coordination.

The first measure would give external evaluators ongoing access to AI companies, allowing them to verify safety practices, report incidents, and assess not only completed models but also training processes. Amodei says Anthropic is committing to this approach and urges other frontier AI companies to adopt it.

The second stage would involve common safety standards and limits on unchecked AI progress among companies operating in democratic countries. The third would seek cooperation with other governments, including China.

Amodei acknowledges that slowing US development while China continues advancing could create a national security risk. His proposed solution is therefore not an unconditional global halt, but a system that attempts to preserve the US lead while creating mechanisms for greater safety.

Trump Sees AI Competition as the Priority

Trump’s response takes the opposite approach. In a series of Truth Social posts, he described concerns that AI could destroy humanity as a “HOAX” and questioned why industry leaders would support regulation that could restrict their own businesses.

Trump also argues that AI and data centers could become the largest economic development engine in history, exceeding the economic impact of industries such as oil, gold, diamonds and the internet.

Trump has repeatedly connected AI development with the competition between the United States and China. His argument is that restrictions on American companies could weaken the US position while allowing China to advance.

“The Trump Administration has stopped AI ‘people’ from doing bad, or potentially bad, ‘things,’ like Dario (Anthropic!),” Trump writes, adding that the administration already has “tremendous CRIMINAL and REGULATORY power” over AI companies. He also characterized opposition to AI and data centers as a “SICK conspiracy” and argued that China would be the primary beneficiary of efforts to constrain the US industry.

That position is consistent with the administration’s broader emphasis on maintaining an American lead in AI. White House advisor David Sacks argues that companies concerned about the risks of their own systems can act voluntarily, while US Vice President JD Vance warns that the United States is competing with China in an AI arms race. House Speaker Mike Johnson, meanwhile, argues that emergency regulation could cause the United States to lose ground to China.

Industry Consensus Splits Over How to Move Forward

The call for greater caution is coming from inside the AI industry itself, but opinions are divided. Altman has backed Amodei’s call to “pace the frontier,” arguing that AI development needs to avoid two outcomes: humanity losing control of increasingly capable systems and excessive concentration of power in a single company, laboratory or country.

“AI must always serve people,” Altman says, arguing that alignment and safety techniques need to remain ahead of model capabilities.

Elon Musk, CEO, Tesla and X, has also supported the idea of additional oversight, proposing that leading AI companies allow competitors to test their models before release. His approach would create a form of peer review, with rival companies evaluating systems for potential safety problems rather than allowing AI developers to assess their own work exclusively.

Musk says this would not be a perfect solution, but that independent competitors would have a greater chance of identifying problems.

Jensen Huang, CEO, Nvidia, has taken a different position from Amodei’s call for slower development. While speaking at the All-In Summit, Huang received a phone call from Trump and discussed the issue publicly.

Trump told Huang that AI developers were “playing right in the hands” of those who wanted to prevent the industry from advancing, including China. Huang responded that he agreed with the president.

The divide illustrates a broader question confronting the AI industry: whether the primary risk is moving too slowly in a geopolitical race or moving too quickly without sufficient safeguards.

For Amodei, the answer is to create mechanisms that make faster development more accountable rather than simply stopping progress. His proposal includes stronger testing, interpretability research, operational security, and external evaluation, alongside international coordination.

The challenge is that the proposed safeguards themselves depend on cooperation among companies and governments that have powerful incentives to maintain their technological advantage.