Anthropic CEO Refuses Pentagon’s Demands Amid Rising Tensions
Dario Amodei, the CEO of Anthropic, asserted on Thursday that the artificial intelligence firm “cannot in good conscience acquiesce” to Pentagon demands for unfettered deployment of its technology.
This declaration marks a significant escalation in an unusually public standoff with the Trump administration, with the potential to jeopardize the company’s government contract as early as Friday.
Anthropic, the creator of the AI chatbot Claude, has indicated a willingness to engage in negotiations. However, it cautioned that recent modifications to contract language proposed by the Defense Department “made virtually no progress” toward preventing the technology’s use for mass surveillance of American citizens or in the development of fully autonomous weapons.
Pentagon spokesperson Sean Parnell countered these assertions on social media, asserting that the military “has no interest in employing AI for mass surveillance of Americans (which is illegal)” nor does it aim to develop autonomous weapons operating independently of human oversight.
Internally, Anthropic has policies that prohibit such applications. The company currently stands out as the sole prominent AI developer—alongside competitors like Google, OpenAI, and Elon Musk’s xAI—that has not agreed to supply its technology to a newly established internal U.S. military network.
Amodei emphasized the value Anthropic’s technology brings to the military, stating, “It is the Department’s prerogative to select contractors most aligned with their vision. However, we sincerely hope they reevaluate their decision.”
The conflict heightened when Defense Secretary Pete Hegseth issued an ultimatum on Tuesday following a meeting with Amodei: permit unrestricted military utilization of Anthropic’s AI technology by Friday or risk losing the Pentagon contract altogether.
Officials further warned that the company could be categorized as a supply-chain risk, or the Cold War-era Defense Production Act could be invoked to extend military control over its products.
Amodei criticized these threats as paradoxical, arguing, “those latter two threats are inherently contradictory: one designates us a security risk; the other labels Claude as indispensable to national security.”
Parnell reiterated the Pentagon’s intent to “leverage Anthropic’s model for all lawful purposes,” though he did not clarify what those purposes might entail. He contended that broader access to the technology is essential to prevent “compromising critical military operations.”
“We will not allow ANY company to dictate the terms of our operational decision-making,” he proclaimed.
Negotiations between Anthropic and Pentagon officials have been underway for several months. Amodei indicated that should the Department of Defense remain intransigent, Anthropic “will work to facilitate a smooth transition to another provider.”
The public dimensions of this dispute have incited criticism from lawmakers. Republican Senator Thom Tillis of North Carolina condemned the Pentagon’s handling of the situation as unprofessional, suggesting that Anthropic is “merely attempting to safeguard us from ourselves.”
“Why are we having this dialogue in public?” Tillis questioned reporters. “This is not how one engages with a strategic vendor holding contracts.” He advocated for a more private discourse, wherein concerns can be properly addressed.
Senator Mark Warner of Virginia, the leading Democrat on the Senate Intelligence Committee, expressed his profound concern over allegations that the Pentagon was “attempting to intimidate a leading U.S. company.”
“Regrettably, this situation is indicative of the Department of Defense’s inclination to disregard AI governance,” Warner stated.
He argued that it “highlights the urgent necessity for Congress to implement robust, binding AI governance mechanisms in national security contexts.”
Pentagon representatives maintain that AI systems will be utilized within the confines of legality, even as the department endeavors to overhaul its internal legal culture.
Hegseth previously articulated a desire for legal advisors who provide constitutional counsel without becoming “impediments.”
His administration has seen the unexplained dismissal of the Army and Air Force’s top legal officers, with the Navy’s chief lawyer resigning shortly after the 2024 election.
Source link: Timesofindia.indiatimes.com.
