Key Takeaways
- President Trump delayed signing an AI executive order because he feared it could weaken the U.S. lead in artificial intelligence.
- The draft order would have created a voluntary framework for the government to review national‑security risks of the most advanced AI systems before public release, involving companies such as Anthropic, OpenAI, and Google.
- Banking regulators and Treasury Secretary Scott Bessent warned Wall Street CEOs about cybersecurity threats posed by powerful models like Anthropic’s Claude Mythos.
- Some administration officials advocated giving trusted cybersecurity experts early access to these AI tools to better defend critical infrastructure.
- Trump’s pro‑AI agenda clashes with voter concerns over jobs, electricity costs, and societal impact, creating internal Republican splits.
- A public legal battle with Anthropic—triggered after the Pentagon barred use of its Claude chatbot—further complicates cooperation.
- Experts note healthy tension within the White House: a desire to test frontier AI for safety versus worries that regulation could slow innovation.
- Vice President JD Vance emphasized the administration’s goal to win the global AI race while protecting citizens from downsides.
- Although the Commerce Department had announced screening agreements with Google, Microsoft, and xAI (later removed from its website), similar reviews already exist under Biden‑era pacts with Anthropic and OpenAI.
Trump Postpones AI Executive Order Over Competitiveness Concerns
President Donald Trump called off a planned White House ceremony to sign a new executive order on artificial intelligence just hours before it was set to occur. He explained that he was worried the measure could dull America’s edge in AI technology, stating, “We’re leading China, we’re leading everybody, and I don’t want to do anything that’s going to get in the way of that lead.” The decision came after Trump reviewed the order’s text and found elements he disliked, prompting him to delay the Oval Office event with tech industry executives.
Proposed Order Would Have Created a Voluntary AI Vetting Framework
According to a source familiar with the White House deliberations, the executive order would have established a framework for the government to vet the national‑security risks of the most advanced AI systems before their public release. The directive was described as a voluntary collaboration with participating U.S.-based tech companies, including Anthropic, OpenAI, and Google. The goal was to subject frontier models to security screening while preserving industry cooperation, though the exact mechanics remained undisclosed.
Banking Industry Raises Cybersecurity Alarms Prompting Treasury Meeting
The push for government review of leading AI systems followed mounting concern within the banking sector and other institutions about AI’s growing ability to uncover cybersecurity vulnerabilities in global software. In April, Treasury Secretary Scott Bessent and outgoing Federal Reserve Chair Jerome Powell convened an urgent meeting with Wall Street CEOs at the Treasury Department’s headquarters. Bessent warned that Anthropic’s AI model, Claude Mythos, was “very powerful” and emphasized the need to convene banks to discuss best practices and future directions for managing such risks.
Administration Considers Sharing AI Tools with Trusted Cybersecurity Experts
In response to the banking sector’s anxieties, some allies of the Republican president proposed better methods for getting powerful AI tools into the hands of trusted cybersecurity experts. The idea was to let those experts evaluate and defend against potential threats posed by advanced models before they become widely available. This approach aimed to strengthen national defenses while still allowing the private sector to innovate, reflecting a broader debate over how best to balance security with technological progress.
Trump’s AI Ambitions Collide with Voter Fears and Partisan Divisions
Trump has long positioned the AI sector as an engine for economic growth, promoting its major players at White House events and international summits, including a recent tech‑CEO gathering alongside China’s Xi Jinping. However, his ambitions have collided with voter apprehensions about AI’s impact on jobs, electricity bills, and everyday life. Republicans themselves are split: some favor embracing the AI industry, while others side with constituents who express skepticism, creating a palpable tension within the administration’s policy discourse.
Legal Clash with Anthropic Fuels Internal Fractures
Cooperation with Anthropic is further complicated by an ongoing legal fight. In February, Trump ordered all U.S. agencies to stop using the company’s chatbot Claude after a highly public dispute between the Pentagon and CEO Dario Anthropic. This ban underscores a broader mistrust that has surfaced within the administration. As noted by Serena Booth, a computer science professor at Brown University and former AI policy fellow, the public back‑and‑forth—“We will release an executive order. No, we won’t. We’re going to sign it this afternoon. Oh, the signing is canceled.”—reflects deeper fractures over how to handle frontier AI.
Experts Highlight the Tension Between Safety and Innovation
Booth observed that the administration is grappling with a “reasonable idea” to test the most capable AI models before release, but fears that excessive government scrutiny could burden developers and slow innovation. “It does come at a potential very large cost to innovation and speed of development,” she said, acknowledging real risks on both sides. Dean Ball, former White House tech policy adviser and now at the Foundation for American Innovation, characterized the disagreements as “healthy tension” in an administration wary of regulating frontier AI firms. He welcomed an executive order that would pull companies closer to the government on cybersecurity but said he is “fine with them taking time to get this right.”
White House Stresses Balance Between Innovation and Protection
At a Tuesday press briefing, Vice President JD Vance declined to discuss specifics of the stalled order but reiterated the administration’s dual objectives: promoting innovation while addressing cybersecurity threats and data privacy. “The president wants us to be pro‑innovation. He wants us to win the AI race against all other countries in the world,” Vance said, adding, “We also want to make sure that we’re protecting people.” He characterized the approach as collaborative with tech firms, acknowledging downsides and emphasizing the need to balance safety against innovation.
Existing Screening Agreements Hint at Ongoing Government‑Industry Cooperation
Despite the halted executive order, similar screening activities are already underway. Earlier this month, the Commerce Department announced agreements with Google, Microsoft, and Elon Musk’s xAI to evaluate their most powerful AI models before public release, building on prior Biden‑era pacts with Anthropic and OpenAI. The announcement later vanished from the department’s website, raising questions about transparency. Nonetheless, these arrangements indicate that the government continues to pursue voluntary, industry‑led reviews of frontier AI, even as the broader executive‑order initiative remains on hold.