Pentagon Officials Push for AI Control from Anthropic
On Wednesday evening, Pentagon officials submitted their final proposal to Anthropic, a leading artificial intelligence startup, amidst ongoing negotiations concerning the military’s use of the company’s AI technology. Sources familiar with the discussions indicate that the offer may not differ significantly from the initial requests being made by the government.
### Deadline for Agreement Set
Defense Secretary Pete Hegseth has established a deadline of Friday evening for Anthropic to grant the military unrestricted access to its AI technology. Failure to comply could result in the cessation of all business relationships with the U.S. Department of Defense. Insiders suggest that the government’s proposal aims to secure comprehensive access to Anthropic’s AI model for military operations.
A senior Pentagon official highlighted that should Anthropic refuse the terms, the company would not only forfeit government contracts but also risk being labeled a supply chain liability. This designation could complicate future partnerships within defense-related sectors.
### Potential Use of Defense Production Act
In light of the stalled negotiations, Pentagon officials are contemplating the invocation of the Defense Production Act as a means of compelling Anthropic to comply with the government’s terms. This legislation could enable the military to exercise greater influence over the distribution and control of essential technologies connected to national security.
Anthropic was awarded a significant $200 million contract in July, aimed at developing innovative AI capabilities that would bolster U.S. military operations and enhance national security. However, as discussions continue, the startup has expressed concerns over the implications of unregulated military AI use.
### Concerns Over Civil Liberties
Anthropic has consistently requested that the Pentagon agree to stipulations aimed at preventing the misuse of its AI model, known as Claude. Notably, the company seeks assurances that Claude will not be used for mass surveillance of American citizens, which is deemed illegal by former Trump administration officials and is also aligned with Pentagon legal policy.
In negotiations, Anthropic’s CEO, Dario Amodei, has emphasized the importance of maintaining human oversight in military operations. He has articulated reservations regarding the reliability of Claude, cautioning against the use of AI for critical targeting decisions without human intervention. Insiders have pointed out that the model is not infallible, and could lead to severe consequences, including unintended escalations or failures in mission objectives.
### Ongoing Negotiations
During a meeting at the Pentagon on Tuesday morning, Defense Secretary Hegseth reiterated the urgent need for Anthropic to finalize an agreement that would offer the military full access to its AI model by the end of the week. While the company’s representatives had not responded to requests for comments as of Thursday morning, the urgency of the situation denotes the critical nature of the ongoing discussions.
As negotiations draw nearer to their conclusion, the outcome will heavily impact not only Anthropic’s business relationship with the Department of Defense but also the broader framework for AI technology’s role in military operations. The tension between innovation and ethical considerations remains at the forefront of this dialogue, as stakeholders attempt to navigate the complexities of integrating AI into national security frameworks responsibly.
Source: Original Reporting
About The Author
