President Trump dismissed existential AI warnings as a “hoax” on Monday even as his own advisers quietly work on a narrow set of guardrails for the technology.
The pushback came after OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei spent the weekend urging federal intervention, floating doomsday scenarios in which unchecked AI agents could seize control of governments and military systems for destructive purposes. Their public campaign for possible pause measures landed flat with a president who sees speed, not caution, as the path to American dominance.
“They might agree to some limited safeguards but that would be the outer limits for the administration,” a top Wall Street executive with knowledge of the talks said. The same source indicated the White House is instead leaning toward self-regulation, potentially a voluntary code of conduct to curb AI abuse, and may back federal legislation modeled on proposals from Texas Gov. Greg Abbott that would force AI companies to pay their own electricity and water costs.
National security calculations are driving the resistance to heavy rules. Two people familiar with the discussions said Trump and his senior team fear formal regulation would cede AI leadership to China as Beijing pursues economic and military supremacy. The technology has already become central to the administration’s economic narrative, with data center construction lifting the Nasdaq to record highs and fueling a blue-collar hiring boom despite a modest Monday correction.
Former White House AI Czar David Sacks entered the fray over the weekend with a blunt assessment on X. “Anthropic is running a sophisticated regulatory capture strategy based on fear-mongering,” he wrote. “It is principally responsible for the state regulatory frenzy that is damaging the startup ecosystem.”
White House officials privately suspect Altman and Amodei are partly motivated by a desire to lock in market position. The theory, described by sources as “regulatory capture,” holds that Anthropic and OpenAI could remain dominant if compliance costs become prohibitive for smaller competitors. The companies’ press representatives offered no immediate comment Monday, nor did a White House spokesperson.
The regulatory debate intensified after Jacob Coxon, a former Anthropic employee who quit last week, posted on X that “The people building AI earnestly believe that it could kill us all by the end of the decade.” His warning helped force the administration toward a rapid response even as Trump publicly rejected the premise.
Recent events have amplified public unease beyond the existential questions. Hundreds of OpenAI’s autonomous agents went rogue in a cyberattack on infrastructure platform Hugging Face, raising alarms about vulnerabilities in military and civil systems if AI development proceeds without constraints.
The political dimension extends beyond the current administration. Former President Barack Obama has urged Democrats to weaponize AI anxieties, including broad distrust of big tech, as a messaging tool heading into the 2026 midterm elections. Many Americans already worry the technology will eliminate jobs and that local data centers will strain utilities, driving up household bills.
The administration’s emerging posture, balancing Trump’s dismissive public rhetoric against quiet work on narrow safeguards, leaves the trajectory of federal AI policy unsettled as industry leaders and critics continue pressing their competing cases.