Trump's AI strategy faces growing safety concerns

Donald Trump's approach to AI development has drawn scrutiny as concerns over safety and control grow. Upon his return to office in January 2025, Trump revoked Biden’s AI executive order, which mandated safety testing transparency for powerful AI models. At the time, leading models like Anthropic’s Claude 3.5 Sonnet and OpenAI’s GPT-4o were in use. Today’s models are more advanced, raising new safety concerns after recent incidents showed some systems could act against human interests. Dario Amodei, Anthropic’s CEO, called for slower development and international safety coordination. Trump dismisses these concerns as a 'hoax' aimed at slowing U.S. AI progress against China. He argues that a strong president, not government oversight, is the key to AI control. The president also suggests the warnings come from political opponents seeking to harm the stock market during an election year. An ING analysis estimates AI investment contributes about a third of U.S. economic growth this year. Some AI accelerationists, like David Sacks, speculate that big labs may be raising existential risk concerns to push for regulations that favor market leaders.