President Trump's abrupt decision to halt a landmark AI safety order reveals a deep conflict within his administration: how to regulate a technology seen as critical to U.S. competitiveness without stifling the very innovation that provides its edge.
President Donald Trump on Thursday scrapped the signing of a major executive order that would have created a voluntary government review process for new AI models, saying he feared the regulation could hand a competitive advantage to China and other rivals.
“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,” Trump told reporters in the Oval Office. “I really thought [the order] could have been a blocker.”
The now-shelved order would have formalized a voluntary framework for leading AI companies like Anthropic and Google to allow agencies such as the National Institute of Standards and Technology (NIST) to vet frontier models for national security risks before public release, according to people familiar with the draft. This move follows Trump’s day-one repeal of a Biden-era order that had mandated leading AI labs to share safety testing results with the government.
The delay is a short-term victory for tech giants, but it prolongs regulatory uncertainty for a multi-trillion dollar industry. The debate now intensifies over whether the U.S. will pursue a "light-touch" approach, favored by some American tech companies, or adopt more stringent rules like the EU’s AI Act, with billions in investment hanging in the balance.
The administration's stance on AI mirrors a broader push to relax technology safeguards across other sectors. In healthcare, proposed rules from the Department of Health and Human Services aim to remove requirements for user-testing and transparency in AI tools used in electronic health records, a move that supporters argue will spur innovation by reducing burdens on developers.
However, patient-safety advocates and groups like the American Hospital Association have raised alarms. The AHA has warned that the "black box nature of certain AI tools" is a growing challenge, and that a lack of transparency could "undermine clinician trust" and erode patient safety.
The debate is not merely academic. At Kaiser Permanente, some psychotherapists using Abridge's new AI-powered scribe software to summarize patient visits say they must constantly correct the notes. Paul Boyer, a therapist, told KFF Health News the tool is "not good at picking up on clinical nuance," a flaw that highlights the gap between AI's potential and its current reliability in critical fields.
The push for pre-deployment testing gained urgency after AI company Anthropic’s Mythos Preview model demonstrated the ability to autonomously discover thousands of critical cyber vulnerabilities in major operating systems. While Anthropic has kept the model private, sharing it only with select partners to bolster defenses, Vice President JD Vance said Wednesday that a "bad actor could use Mythos to target various cybersecurity vulnerabilities."
For investors in AI leaders like Microsoft, Google, and Anthropic, the regulatory landscape remains a key risk factor. While the canceled order removes an immediate hurdle, the lack of a clear federal framework in the U.S. contrasts with the EU's comprehensive AI Act. This regulatory divergence could create compliance challenges for companies operating globally and influence long-term capital allocation in a sector where a single large training run can cost billions of dollars.
This article is for informational purposes only and does not constitute investment advice.
