Key Points:
- The Trump administration sent conflicting messages about a potential executive order that could require companies to obtain government approval for new software.
- Fears that Anthropic would release its new Mythos model prompted officials to reconsider their original hands-off approach to technology regulation.
- While 1 official suggested strict rules similar to prescription drug approvals, another official tweeted that the administration wants to avoid heavy bureaucracy.
- Exactly 5 major artificial intelligence labs agreed to participate in voluntary testing, pushing back against the idea of mandatory government reviews.
The Trump administration is sending mixed signals to the technology industry regarding artificial intelligence regulation. At least 7 different lobbyists and policy advisers report struggling to get clear answers about a potential new executive order. This order could force technology companies to undergo strict government vetting before releasing new artificial intelligence models. Right now, the lack of clarity leaves industry insiders scrambling to understand the actual plan.
One senior tech lobbyist explained that different factions inside the White House hold completely opposing views on the issue. This lobbyist noted that poor organization increases anxiety across the tech sector and stops officials from developing effective policies. The administration regularly fails to invite the right experts into the room to discuss these complex topics. A White House official simply called the rumors about the executive order pure speculation and stated that President Donald Trump will announce any official policy directly.
Many industry leaders initially expected Trump to take a completely hands-off approach to artificial intelligence. That assumption broke down this spring when Anthropic released a powerful new model called Mythos. Experts warn that Mythos can supercharge severe cybersecurity threats. Top Trump officials, including Chief of Staff Susie Wiles and National Cyber Director Sean Cairncross, reacted quickly to the threat. They started discussing several possible actions, including 1 strict executive order requiring companies to secure a green light from the government before launching new software.
The messaging from the White House then turned chaotic. On Wednesday, National Economic Council Director Kevin Hassett told reporters that the administration considered using a strict approval process similar to that used by the Food and Drug Administration to approve prescription drugs. His comments immediately stirred up intense anxiety across the tech sector. To calm the nerves, Wiles posted a message on social media the very next day. She insisted that the administration is pursuing a strategy that empowers American innovators rather than creating heavy bureaucracy.
A second lobbyist mentioned that Wiles offered some comfort with her message, but the overall goals of the White House remain a mystery. Meanwhile, the technology lobby maintains a very consistent message for the new administration. Exactly 3 different lobbyists said tech companies gladly welcome the government’s examination of their new models. However, they insist that any vetting program must remain strictly voluntary.
The industry wants the Center for AI Standards and Innovation to handle these reviews. This office operates under the National Institute of Standards and Technology. Most major American artificial intelligence labs have already agreed to submit their models for this voluntary scrutiny. Exactly 5 leading companies, including Anthropic, OpenAI, Google DeepMind, xAI, and Microsoft, recently announced new partnerships to support this exact review process. Lobbyists view this voluntary step as the best option and argue that the government should issue formal regulations only if the voluntary system fails.
Daniel Castro, president of the Information Technology and Innovation Foundation, pointed out that the messy communication strategy causes unnecessary panic. His think tank receives funding from major players like Anthropic, Microsoft, and Meta. Castro suggested the administration should give people time to understand proposals and provide feedback before floating random ideas to the press.
Companies feel extremely frustrated by the constantly changing timelines. One policy adviser mentioned that clients went through the entire Pentagon procurement process and still received almost no concrete guidance. Tech leaders recently reached out directly to Office of Science and Technology Policy Director Michael Kratsios, Treasury Secretary Scott Bessent, and Cairncross. They used these meetings to argue against mandatory pre-approval rules and ask for a solid timeline regarding the executive order.
Democrats also voiced their frustrations with the slow, uncoordinated response. Senate Minority Leader Chuck Schumer wrote a letter to Homeland Security Secretary Markwayne Mullin on Friday. Schumer expressed deep concern that the administration still has not helped state and local governments prepare for advanced cyberattacks driven by tools like Mythos. He criticized the administration for lacking an effective plan to coordinate with tribal and territorial leaders.
Right now, the competing White House factions sit at a standstill. Lobbyists believe the internal arguments will continue until Trump makes a final determination on the executive order. Even if the president ultimately rejects mandatory testing, industry insiders know the fight will continue. As models grow smarter, or if Democrats take control of the government in the future, mandatory rules will likely become a reality. To prepare for this inevitable shift, many companies are planning to participate in voluntary testing now to stay ahead of future laws.
