2 months ago
34
ARTICLE AD BOX
Anthropic CEO Dario Amodei is reportedly seeking to reopen talks with the US Department of War aft negotiations implicit entree to the company’s AI models broke down past week. The improvement comes days aft Amodei criticised OpenAI’s defence woody and called CEO Sam Altman’s messaging “straight up lies” successful an interior memo to employees.A study by The Financial Times cited aggregate radical with cognition of the substance to assertion that Amodei has been successful discussions with Emil Michael, the Pentagon’s under-secretary of defence for probe and engineering, to scope a caller declaration that would let the US subject to proceed utilizing Anthropic’s technology. The talks travel warnings from Defence Secretary Pete Hegseth that Anthropic could beryllium designated a proviso concatenation hazard if an statement is not reached.
The quality began aft Anthropic declined the Pentagon’s petition for unrestricted entree to its AI systems, citing concerns astir safeguards astir home wide surveillance and autonomous weapons. The Pentagon aboriginal signed a woody with OpenAI instead, prompting Amodei to criticise the statement and reason that his institution rejected the presumption due to the fact that “they cared astir placating employees, and we really cared astir preventing abuses.”
Why did talks betwixt the Pentagon and Anthropic interruption down
Last week, tensions escalated betwixt Anthropic and the US Department of War aft negotiations implicit the usage of the company’s AI exertion broke down.
Emil Michael, the Pentagon’s under-secretary of defence for probe and engineering, reportedly described Anthropic CEO Dario Amodei arsenic a “liar” with a “God complex”.Talks collapsed a time aboriginal aft some sides failed to hold connected declaration connection that Anthropic said was indispensable to forestall its AI systems from being utilized for wide home surveillance. The institution has besides said it opposes the usage of its exertion for lethal autonomous weapons.“Near the extremity of the dialog the [department] offered to judge our existent presumption if we deleted a circumstantial operation astir ‘analysis of bulk acquired data’ which was the azygous enactment successful the declaration that precisely matched this script we were astir disquieted about. We recovered that precise suspicious,” Amodei wrote successful a memo to employees connected Friday, archetypal reported by The Information and aboriginal seen by the Financial Times.In the memo, Amodei besides wrote that messaging from the Pentagon and OpenAI, which signed its ain statement with Defence Secretary Pete Hegseth, was “just consecutive up lies astir these issues oregon tries to confuse them”.He further suggested that Anthropic whitethorn person been sidelined due to the fact that “we haven’t fixed dictator-style praise to Trump,” drafting a opposition with OpenAI's chief, Sam Altman.Anthropic was archetypal awarded a $200 cardinal declaration with the US Department of War successful July 2025 and became 1 of the aboriginal AI exemplary providers utilized successful classified settings and by nationalist information agencies.The quality intensified aft the Pentagon pushed for AI companies to let their exertion to beryllium utilized for immoderate “lawful” purpose. The concern culminated past week erstwhile Hegseth said helium planned to designate Anthropic a proviso concatenation risk, a measurement that would necessitate companies successful the subject proviso concatenation to chopped ties with the AI firm.
