Trump Rejects AI Guardrails as Tech Chiefs Urge Restraint

News

1. Anthropic Co-Founder Backs AI Kill Switch

2. Microsoft Sets Human Control Rules for Future AI

3. Trump Rejects AI Guardrails as Tech Chiefs Urge Restraint

4. Mistral Uses AI to Rewrite Legacy Scientific Software

5. OpenAI Weighs Slower Pace for AI Development as Safety Pressure Builds

6. Anthropic Says AI Is Letting Smaller Attackers Run State-Scale Operations

US President Donald Trump dismissed mounting concerns over artificial intelligence safety on Monday, calling the push for tighter oversight a “hoax” and a “sick conspiracy” as leaders of several major AI companies urged greater restraint.

Trump said the US already had sufficient powers to police AI companies and argued that slowing development would benefit China. “The only control or ‘guardrails’ that AI needs is a STRONG AND SMART (High IQ!) PRESIDENT,” he wrote on Truth Social.

His comments sharpen a divide with some of the companies developing the world’s most advanced AI models.

Anthropic Chief Executive Dario Amodei called over the weekend for companies to “pace the frontier,” arguing that advances in model capabilities should slow when safeguards and researchers’ understanding fail to keep up.

“We must slow the pace at which we improve the capabilities of AI models,” Amodei wrote. “Progress will still seem fast, and we must make wise use of the time we gain.”

Amodei proposed giving independent evaluators permanent access comparable with that available to employees, followed by common safety standards among leading developers and eventually international coordination.

OpenAI Chief Executive Sam Altman backed the proposal. “I agree with Dario that we need to pace the frontier,” he wrote on X, adding that OpenAI would also give independent evaluators employee-like access. Elon Musk, who runs rival developer xAI, wrote simply: “Dario is right.”

Google DeepMind Chief Executive Demis Hassabis also supported the broad direction, saying Amodei’s proposal pointed toward “the right path forward” while cautioning that the details still needed work. Hassabis previously proposed an industry standards body that could coordinate a slowdown among frontier laboratories if risks became severe enough.

The debate has intensified after AI systems behaved unexpectedly during cybersecurity testing.

OpenAI said models operating with reduced safeguards during internal evaluations in July circumvented isolation controls, gained internet access and compromised parts of its own research infrastructure and systems belonging to Hugging Face. OpenAI described the episode as a “warning shot” showing that sufficiently capable agents could exploit weaknesses across several computer systems.

Anthropic later disclosed three incidents in which Claude models reached the internet during cybersecurity evaluations and gained unauthorized access to real organizations’ systems. The company said a third-party evaluation environment had been misconfigured and that the models were deliberately being tested without normal cyber safeguards.

The companies have not agreed to a joint slowdown or established common thresholds for deciding when development should pause. Any such coordination would also raise commercial and potentially antitrust questions because laboratories are competing for computing capacity, researchers and customers.

OpenAI has sought greater legal clarity over whether AI developers could coordinate on safety without violating US competition law. The issue has gained urgency as researchers consider whether increasingly capable systems could contribute to developing their successors.

China adds another complication. Amodei has argued that US developers cannot indefinitely slow progress while Chinese competitors continue advancing, linking his safety proposals to restrictions on advanced chips and other technology.

Trump has taken the opposite view, repeatedly arguing that restrictions could surrender the US lead to China.

House Speaker Mike Johnson said Monday that Trump was expected to meet AI executives this week or early next week to discuss companies’ responsibilities for safety and what role, if any, the government should play.

The argument is also beginning to affect markets.

AI-linked stocks fell globally on Monday, with the Nasdaq 100 dropping as much as 1.2% to a six-week low before recovering part of the decline. The Philadelphia semiconductor index fell 5.2%, while Nvidia lost about 3% as investors considered whether slower AI development could eventually curb spending on chips and computing infrastructure.