OpenAI is modifying its agreement with the Pentagon to guarantee security measures for surveillance: Sam Altman, CEO

OpenAI has said it will amend its newly signed contract with the US Department of Defense to ensure its artificial intelligence models are not used for domestic surveillance, following criticism that the deal gave too much unchecked power to military officials.

Chief executive Sam Altman announced on X that, consistent with applicable laws including the Fourth Amendment of the US Constitution, OpenAI’s systems “shall not be intentionally used for domestic surveillance of US persons and nationals.” He added that the agreement would include clearer safeguards.

The defence contract was secured on Friday, shortly after rival AI firm Anthropic declined to allow unconditional military use of its Claude models. Anthropic had sought guarantees that its AI would not be deployed for surveillance or autonomous weapons. According to Altman, the Defense Department agreed to introduce technical safeguards addressing those concerns in OpenAI’s case.

However, questions were raised about whether such safeguards would be sufficient or enforceable. In response, Altman acknowledged that the rollout of the announcement had been rushed and said the company would revise the agreement to make its principles clearer. He also stated that OpenAI’s models would not be used by US intelligence agencies, including the National Security Agency.

Altman admitted that the issues surrounding military use of AI are highly complex and require careful communication. He noted that there are areas where the technology is not yet mature and where safety trade-offs are not fully understood.

At the same time, he criticised the Pentagon’s move to designate Anthropic as a supply chain risk after it refused unconditional terms. The designation effectively bars companies working with the US military from engaging with Anthropic. The company has said it will challenge the order in court, describing it as a dangerous precedent for American businesses negotiating with the government.

The developments highlight growing tensions over how advanced AI systems should be deployed in defence and intelligence operations, and the balance between national security and civil liberties.