A senior Pentagon official refuted Anthropic researchers' claims that "AI will eliminate humanity": It's just fear of change; don't fall into doomsday scenarios.

The rift between the Pentagon and Anthropic continues to widen. The U.S. Department of Defense's chief technology officer publicly refuted recent widely discussed warnings about the risks of AI, while also disclosing that the Department of Defense is accelerating the migration of AI workloads from Anthropic and characterizing the legal dispute between the two parties as a core controversy involving national security.

On Thursday, at a defense industry conference in Washington, D.C., U.S. Defense Chief Technology Officer Emil Michael directly responded to a post by former Anthropic engineer Jacob Coxon warning about AI, which has now been viewed over 150 million times. The post warned that AI insiders believe the technology "could kill us all before the end of this decade."

Emil Michael dismissed these concerns as a "doomsday cycle," arguing that they were essentially nothing more than "fear of data centers and fear of change."

Meanwhile, Emil Michael revealed that the Department of Defense has migrated approximately 90% of its classified AI workloads from Anthropic to other vendors, with the remainder expected to be transferred by the end of this month.

On the legal front, the litigation between the two sides is still ongoing—Emil Michael has characterized the pending case in the District of Columbia Circuit as a "more important" battleground and insists that Anthropic still poses a risk to the U.S. supply chain in the absence of a preliminary injunction.

"Doomsday tweets" spark controversy; Pentagon officials publicly refute them.

Earlier this week, Coxon warned in a series of posts that OpenAI and Anthropic are "heading straight toward self-improving superintelligence, gambling with our lives," and that "superhuman systems capable of hacking into any system, disrupting any field overnight, and gaining real power and resources will soon emerge."

The above content quickly spread virally on social media.

Emil Michael directly addressed the post at the meeting, urging it to be de-escalated. "It's easy to fall into this doomsday cycle—'We're going to lose 40% of Americans, we're going to destroy humanity,'" he said. "It's just condensing all the fear into a well-written tweet."

Emil Michael also stated that the potential harms of AI can be mitigated through free market mechanisms, industry collaboration, and active interaction with the government. He previously served as a senior executive at Uber Technologies Inc. and currently leads the push for the large-scale application of AI technology in the U.S. military system.

Procurement dispute escalates into court battle

The conflict between the Pentagon and Anthropic stems from the latter's additional security requirements for AI tools intended for military use.

Emil Michael led negotiations with the company, but the talks broke down earlier this year when the Department of Defense rejected Anthropic's terms, deeming them an infringement on military authority, and declared the company a threat to the U.S. supply chain.

The dispute then moved to federal court. Anthropic filed lawsuits in California and Washington, D.C., seeking to overturn the Department of Defense's findings. The California court has ruled in favor of Anthropic, but the U.S. Court of Appeals for the District of Columbia has yet to issue a decision.

Emil Michael accused Anthropic of "choosing a jurisdiction that is favorable to him" in the politically liberal Northern District of California, and characterized the case in the District of Columbia Circuit as "more important" and "more national security oriented," expecting a ruling in the coming months.

Workloads are being moved out at an accelerated pace, and the Ministry of National Defense is drawing clear lines.

In terms of practical operation, the Department of Defense has already migrated about 90% of its classified AI workloads from Anthropic, and the remaining portion is expected to be fully transferred by the end of this month.

Emil Michael told reporters after the meeting that "all the work related to Maven Smart Systems or Palantir was completed several months ago," referring to the AI system of the well-known defense technology company Palantir.

This statement contrasts sharply with previous signals from Commerce Secretary Howard Lutnick, who had indicated that Anthropic had reconciled with the government after a "minor incident."

But Emil Michael clearly distinguished the Department of Defense’s national security needs from the positions of other federal agencies.

"I'm glad they've resolved the issues with the rest of the government," he said. "But the Department of Defense's position has always been very clear: the way we need to work with AI vendors must serve combatants and national defense. This means that if we're going to use software, it must be able to be used for all legitimate purposes without constraint."

Risk Warning and DisclaimerInvesting involves risk; please exercise caution. This article does not constitute personal investment advice and does not take into account the specific investment objectives, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Any investment decisions made based on this information are at your own risk.