OpenAI CEO Sam Altman has defended the company’s decision to allow ChatGPT to be used for classified assignments by the US government, as per multiple media reports. The ChatGPT maker had faced a lot of criticism after it announced the deal with the Pentagon last week just hours after US Defense Secretary Pete Hegseth said that Anthropic would be categorized as a supply chain risk, a designation usually reserved for adversarial foreign companies.
Altman said, as quoted by the Wall Street Journal, that he didn’t regret signing the deal with the Defense Department but he wished that he hadn’t announced the decision so quickly, making it look “opportunistic” and “not united with the field.”
Altman tells employees, ‘You don’t get to weigh in’
Meanwhile, as per a CNBC report, Altman also told the staff that OpenAI does not “get to make operational decisions” on how its technology is used by the Department of Defense.
“So maybe you think the Iran strike was good and the Venezuela invasion was bad,” Altman said. “You don’t get to weigh in on that.”
As per a Bloomberg report, Altman also suggested that the desire to regulate how the DoD uses its AI may have been part of the tensions between the Pentagon and Anthropic.
Reportedly, the US government had used Anthropic’s AI during the capture of Venezuela president Nicolas Maduro and the recent strikes in Iran. Anthropic is also said to have asked questions on how its AI was used in the capture of Maduro, which is said to have upset the Defense Department officials.
OpenAI had signed a $200 million agreement with the Pentagon last year that allowed the agency to use its models in non-classified use cases. The agreement last week also permitted it to deploy the models across classified networks.
However, the Pentagon is also said to have talked with Elon Musk’s xAI to allow for using its models across classified use cases.
OpenAI had set three conditions for the use of its AI models by the Pentagon: no use of its AI for mass domestic surveillance, direct autonomous weapons systems or high-stakes automated decisions.
Altman reportedly noted that xAI will, however, pose no such requests to the Pentagon and will do whatever the agency says.
“I believe we will hopefully have the best models that will encourage the government to be willing to work with us, even if our safety stack annoys them,” Altman was quoted by CNBC as saying. “But there will be at least one other actor, which I assume will be xAI, which effectively will say ‘We’ll do whatever you want.’”