Anthropic CEO Dario Amodei released a statement on Thursday, clarifying that the AI startup will not accept demands of the US Department of Defense when it comes to unrestricted use of its AI models. The Anthropic boss claimed that accepting these conditions could “undermine, rather than defended, democratic values.”
This announcement comes after a reported tussle between Amodei and US Defense Secretary Pete Hegseth.
Dario Amodei rejects Pentagon’s demands for AI use
In the statement, the Anthropic CEO asserted that the company wants AI to be used to protect democratic countries such as the United States. However, Dario Amodei believes that current AI technology could not safely meet the demands of the US Military.
He said, “Some uses are also simply outside the bounds of what today’s technology can safely and reliably do.”
US Defense Under Secretary says Amodei has a god complex
After Amodei’s announcement, US Defense Under Secretary Emil Michael slammed the Anthropic chief. He claimed that Dario Amodei was a “liar” and had a “god complex.” Michael reckoned that Dario just wanted to “personally control the US Military,” by not accepting its demands.
Michael added that despite claims made by Amodei, the US Department of Defense would “always adhere to the law.” However, it was not going to “bend to whims of any one for-profit tech company.”
Which demands did Anthropic reject?
Two main conditions became the centre of this standoff – mass domestic surveillance and the development of autonomous weapons.
Dario said that while AI systems can be used for lawful foreign intelligence and counterintelligence missions, using these systems for mass domestic surveillance was “incompatible with democratic values.”
The Anthropic boss raised concerns that AI could allow authorities to gather “scattered, individually innocuous data into a comprehensive picture of any person’s life—automatically and at massive scale.”
When it comes to the development of autonomous weapons, Amodei insisted that while these could “prove critical” for national defense in the future, AI systems were “simply not reliable enough” to power fully autonomous weapons just yet.
He admitted that allowing for such a use case could put American citizens at risk. Amodei asserted that such systems would need to be deployed with proper guardrails, “which don’t exist today.”
What happens now?
Anthropic reportedly faced a Friday deadline to meet demands of the Pentagon. As per reports, the AI startup could be deemed a “supply chain risk” if it fails to accept the terms. This would prohibit Anthropic from working with the US Defense Department and any defense contractor.
Dario Amodei has stated that Anthropic would be willing to allow for a transition to another AI provider in case it is removed from the Pentagon’s networks. He said, “Should the Department choose to offboard Anthropic, we will work to enable a smooth transition to another provider.”
Axios previously reported that the Pentagon had reached out to defense suppliers, Boeing and Lockheed Martin, to understand their potential reliance on Anthropic’s AI systems.
Anthropic is currently the only company that has its AI systems running on the US’ classified networks. Though companies such as xAI and Google have signed deals with the US government that allow for unrestricted AI use for “all lawful purposes,” as per reports.
