You can mark a year by product launches or by pivotal moments that shift our perception of AI. The AI sector constantly delivers news, from major takeovers, indie triumphs, public pushback against dubious products, to high-stakes contract negotiations. We’re unraveling this complexity to see where we stand and how the year has unfolded so far.
Anthropic vs. the Pentagon
Anthropic CEO Dario Amodei and Defense Secretary Pete Hegseth hit a deadlock in February over contract terms about the U.S. military’s use of Anthropic AI tools.
Anthropic opposed its AI being used for widespread surveillance in America or autonomous weapons. Meanwhile, the Pentagon insisted the Department of Defense, now called the Department of War by the Trump administration, should have unrestricted access to Anthropic’s AI models for all “lawful use.” Government representatives were upset about military use limitations set by a private firm, but Amodei stood firm.
“Anthropic understands that the Department of War, not private companies, makes military decisions. We have never interfered with specific military operations nor limited the use of our technology arbitrarily,” Amodei stated in a statement. “However, in narrow cases, we believe AI might threaten, not defend, democratic values.”
The Pentagon set a deadline for Anthropic to accept their terms. Employees at Google and OpenAI signed a letter urging leaders to uphold Amodei’s boundaries on autonomous weapons and surveillance.
The deadline passed without Anthropic’s compliance. Trump ordered federal agencies to phase out Anthropic tools over six months and labeled the company, valued at $380 billion, a “radical left, woke company” in a social media post. The Pentagon proceeded to declare Anthropic a “supply-chain risk,” typically reserved for foreign threats, barring companies working with Anthropic from military contracts. (Anthropic has challenged this designation.)
Meanwhile, rival OpenAI announced an agreement to deploy its models in classified situations. The tech community was surprised as reports suggested OpenAI would adhere to Anthropic’s military use guidelines.
Techcrunch event
San Francisco, CA | October