AI selects targets in Iran. In the background there is a conflict with the technology supplier
As the daily writes, citing American officials, the Pentagon uses the Maven Smart System (MSS) from Palantir, integrated with Claude. This system, also based on Palantir's massive intelligence collection and open sources, was supposed to suggest “hundreds of targets” in Iran, also ranking them in order of priority. Claude was also to be used to assess strikes, track logistics and compile intelligence information.
What targets does the Pentagon choose in Iran?
Who created the Claude system used by the Pentagon?
What disputes does the Pentagon have with Anthropic?
When did Claude's Maven integration start?
Dispute between Pentagon and Anthropic
While Claude's integration with Maven began during the Joe Biden administration, use of the system has dramatically increased during Donald Trump's administration. Last year 20,000 people were supposed to use it. soldiers. The Washington Post and other media previously reported that Claude was also used during the operation to capture Venezuelan leader Nicolas Maduro.
In recent weeks, the Pentagon has been in a heated dispute with Claude's creator, Anthropic, due to the company's demands that its technology not be used in fully autonomous weapons and for mass surveillance of Americans. In response to this, the head of the Pentagon, Pete Hegseth, announced that he would formally recognize the company as a “threat to the supply chain”, i.e. he would impose on it sanctions applicable to, for example, companies from China.
These sanctions, which have caused a stir in Silicon Valley, are intended to force Anthropica's partners, including: Palantir – to abandon cooperation with it. Experts predicted that this could potentially lead to the bankruptcy of the company. The Pentagon has not yet formally implemented its decision. He also announced that he expects the contract with Anthropik to continue to be implemented over the next six months.
Just after the announcement of Claude's withdrawal from US government systems, Anthropica's rival, OpenAI, concluded an agreement with the Pentagon, agreeing to accept the Pentagon's conditions regarding the use of its technology.
