Anthropic and the Pentagon cannot agree on how AI will be used in the military.
The US Department of Defense has entered into a heated dispute with Anthropic over how the military plans to use new technologies. The sticking point is the company's demands to prevent its systems from being used for autonomous weapon guidance and domestic surveillance.
According to sources familiar with the negotiations, the parties discussed the terms of a large contract worth up to $200 million , but ultimately reached an impasse. Anthropic insists on safeguards that will prevent its models from being used without human intervention to select strike targets or to monitor US citizens. The Defense Department disagrees with this approach.
Company representatives stated that their technology is already widely used by US national security agencies and that dialogue with the military is ongoing. The agency did not immediately respond to journalists' inquiries.
The situation has become one of the first serious tests of the relationship between artificial intelligence developers and the US government. After several years of tension, major tech companies have begun to collaborate more actively with the government, including on military projects. Now, the question of whether private developers can influence how the military and intelligence agencies use such systems in combat is being addressed.
Sources report that the military cites its AI implementation strategy and believes it can use commercial developments regardless of internal company policies, as long as they don't violate US laws. However, deploying their models without the involvement of Anthropic specialists will be difficult, as the company's employees customize and refine the systems for specific tasks.
The conflict comes at a challenging time for Anthropic. The San Francisco-based company is preparing for a possible IPO and is actively promoting its solutions to government customers. It previously won a Pentagon contract along with several other major AI developers.
Anthropic CEO Dario Amodei recently publicly stated that artificial intelligence should aid national defense, but not in scenarios that make the state resemble authoritarian regimes. He also previously spoke out harshly about the deaths of US citizens during protests against immigration policies, which heightened tensions around the use of technology in law enforcement operations.
The US Department of Defense has entered into a heated dispute with Anthropic over how the military plans to use new technologies. The sticking point is the company's demands to prevent its systems from being used for autonomous weapon guidance and domestic surveillance.
According to sources familiar with the negotiations, the parties discussed the terms of a large contract worth up to $200 million , but ultimately reached an impasse. Anthropic insists on safeguards that will prevent its models from being used without human intervention to select strike targets or to monitor US citizens. The Defense Department disagrees with this approach.
Company representatives stated that their technology is already widely used by US national security agencies and that dialogue with the military is ongoing. The agency did not immediately respond to journalists' inquiries.
The situation has become one of the first serious tests of the relationship between artificial intelligence developers and the US government. After several years of tension, major tech companies have begun to collaborate more actively with the government, including on military projects. Now, the question of whether private developers can influence how the military and intelligence agencies use such systems in combat is being addressed.
Sources report that the military cites its AI implementation strategy and believes it can use commercial developments regardless of internal company policies, as long as they don't violate US laws. However, deploying their models without the involvement of Anthropic specialists will be difficult, as the company's employees customize and refine the systems for specific tasks.
The conflict comes at a challenging time for Anthropic. The San Francisco-based company is preparing for a possible IPO and is actively promoting its solutions to government customers. It previously won a Pentagon contract along with several other major AI developers.
Anthropic CEO Dario Amodei recently publicly stated that artificial intelligence should aid national defense, but not in scenarios that make the state resemble authoritarian regimes. He also previously spoke out harshly about the deaths of US citizens during protests against immigration policies, which heightened tensions around the use of technology in law enforcement operations.
