It appears that Anthropic CEO Dario Amodei is sticking to his guns as he insists that his company’s AI models should not be used by the military to conduct mass surveillance of US citizens or control killer drones without human intervention.
The CEO had issued strong criticism in an internal memo last week leaked in InformationFellow OpenAI co-founder and CEO Sam Altman has been accused of taking a knee, claiming he had “praised President Donald Trump in dictator style” and bribed him with donations.
The criticism put Anthropic in a precarious position. The other shoe finally dropped on Thursday, when the Pentagon officially named Anthropic A supply chain risk “effective immediately”, a first for a US company ever been labeled like this. The unprecedented move is especially bizarre given that the Pentagon is reportedly using the cloud to select targets in its war against Iran.
Now, instead of apologizing and moaning in favor of the President, Amodei has spoken in a completely different tone.
“This was a difficult day for the company, and I apologize for the tone of the post,” he wrote in an uncharacteristically brief statement. Thursday’s statement Published on the company’s website.
“This does not reflect my careful or considered views,” Amodei wrote with regret. “This was also written six days ago, and is an outdated assessment of the current situation.”
Amodei also said he regretted that the memo had become public.
“Anthropic did not leak this post nor direct anyone else to do so – it is not in our best interest to escalate this situation,” they wrote.
The AI company has had “productive conversations” with the Defense Department, he said, “about ways we can serve the Department that adhere to our two narrow exceptions, and about ways to ensure a smooth transition for us if that’s not possible.”
Amodei stressed in the note that Anthropic still does not believe that “it is the role of Anthropic or any private company to be involved in operational decision making – that is the role of the military.”
“Our only concerns are our exceptions on fully autonomous weapons and large-scale domestic surveillance, which relate to high-level use areas, not operational decision making,” he wrote.
“Anthropic has a lot more in common with the War Department than we have differences with it,” Amodei offered. “We are both committed to advancing U.S. national security and protecting the American people, and agree on the urgency of implementing AI across the government.”
Amodei’s petition highlights how much pressure his company is under at this time.
But whether MEA’s culpability will be enough to secure a settlement with the Pentagon – especially now that Anthropic has been turned into a red, blinking button that the government doesn’t want any company touching with a ten-foot pole – remains to be seen.
More information on Amodei: Dario Amodei says Trump is a dictator
