Anthropic CEO Dario Amodei has slammed OpenAI’s deal with the US Department of Defense over the use of AI. In an internal memo, Amodei claimed that the statements made by the Sam Altman-led startup were full of lies. Sam Altman tried to position OpenAI as a de-escalater after the Pentagon’s fallout with Anthropic.
According to a report by The Information, Amodei described OpenAI’s characterisation of the Pentagon contract as “straight up lies,” and labelled their approach to safety as “safety theatre.”
OpenAI had reached an agreement with the US Department of Defense just hours after Anthropic’s contract was terminated. Though Sam Altman later admitted that his rushed deal with the Pentagon looked sloppy and opportunistic.
Dario Amodei reveals main reason OpenAI accepted Pentagon deal
In the memo, Amodei claimed that the real reason OpenAI took the Pentagon deal was because the AI firm was not focused on preventing misuse of AI systems, something that Anthropic did care about. He wrote, “The main reason [OpenAI] accepted [the DoD’s deal] and we did not is that they cared about placating employees, and we actually cared about preventing abuses.”
Anthropic had previously stated that it was walking away from the deal over fears of its systems being used in domestic mass surveillance or development of autonomous weaponry. The Pentagon reportedly refused to acknowledge these safeguards and wanted to have the potential to use AI for all “lawful” purposes.
Though OpenAI has stated that it will not allow its systems to be used for such cases either. Sam Altman also shared an amended agreement a couple of days ago, emphasising on these red lines.
Dario Amodei says US government doesn’t like Anthropic because of political donations
In the same memo, Dario Amodei alleged that another reason the US government may not be happy with Anthropic was because the AI firm did not donate to President Donald Trump’s election campaign last year. On the other hand, OpenAI President Greg Brockman and his wife donated $25 million to a Trump super PAC.
Amodei added, “The real reasons [the Department of War] and the Trump admin do not like us is that we haven’t donated to Trump (while OpenAI/Greg have donated a lot).”
What happened to Anthropic’s contract?
Anthropic’s contract was terminated by US President Donald Trump after which the company was designated as a supply chain risk.
Later, US Defense Undersecretary Emil Michael called Amodei a “liar” with a “god complex” over his claims. Michael stated that Anthropic put conditions such as prohibiting the military to use data acquired from public sources like LinkedIn.
But do note that the US is using Anthropic’s AI in its strikes on Iran. The US Government is expected to go through a 6-month transition period as it switches to OpenAI’s models.
Tech group sends letter to Pentagon
The Information Technology Industry Council (ITI), which consists of companies such as Google and Nvidia sent a letter to the Pentagon on Wednesday. The letter does not explicitly name Anthropic, but likely refers to its supply chain risk designation threatened by Pete Hegseth.
The letter states, “We are concerned by recent reports regarding the Department of War’s consideration of imposing a supply chain risk designation in response to a procurement dispute.”
The ITI urged the US Department of Defense to resolve disputes via discussions or to look for alternate providers instead of putting this label on the firm. The letter read, “Emergency authorities such as supply chain risk designations exist for genuine emergencies and are typically reserved for entities that have been designated as foreign adversaries.”
Companies labelled as supply chain risk are prohibited from working with any US defense contractor.
Dario Amodei calls Anthropic public heroes
OpenAI received massive backlash online following its deal with the Pentagon. Users started uninstalling ChatGPT and switching to Anthropic’s Claude chatbot. The OpenAI chatbot saw uninstalls surge by almost 300 per cent.
Claude topped the charts in the US App Store, and even went down on two separate occasions due to surge in usage.
Dario Amodei was aware of this public sentiment. He wrote that while Altman was “presenting himself as a peacemaker and dealmaker,” the public was not accepting his view. Amodei added, “I think this attempted spin/gaslighting is not working very well on the general public or the media, where people mostly see OpenAI’s deal with the DoW as sketchy or suspicious, and see us as the heroes (we’re #2 in the App Store now!).”
However, Amodei was worried that some OpenAI employees may fall for Altman’s messaging and not understand the situation. He added, “It is working on some Twitter morons, which doesn’t matter, but my main worry is how to make sure it doesn’t work on OpenAI employees.”
