Forum Topics Anthropic versus the US DoD (DoW)
Clio
Added a month ago

So this has come to pass. Anthropic held to their ethical standards and will now pay the price. Hegseth has directed the DoW to classify all Anthropic-based software as a "supply chain risk to national security" and Trump proclaimed (loudly on Truth Social) that he is ordering all federal agencies to immediately cease all use of Anthropic technology.

That doesn't just mean that the federal agencies will have to remove all Anthropic technology. It also means that every single supplier to any federal agency will have to remove Anthropic technology from their products supplied to said federal agencies. That's HUGE.

Some are pointing out that the Anthropic situation raises the question of what restrictions all the other AI-firms have with the US government (or anyone else). Specifically, have Microsoft, OpenAI, Grok, etc all agreed to allow the US government (or anyone else) to use their AI to surveil whoever they want and fire weapons autonomously?

A new era indeed. Like an updated version of 1984 combined with Animal Farm (an AI-robots version)

13
Clio
Added a month ago

Just read through the Guardian's report on the above development, and the (theoretical) consequences for Anthropic are actually worse. Hegseth's declaration reads: "Effective immediately, no contractor, supplier, or partner that does business with the US military may conduct any commercial activity with Anthropic."

Anthropic have replied: "We will challenge any supply chain risk designation in court." Adding that applying that designation was "an unprecedented action...never before publicly applied to an American company."

Top executives at other AI companies have publicly sided with Anthropic, including Sam Altman who in an interview on Friday stated that OpenAI shares teh same red lines as Anthropic.

Nearly 500 OpenAI and Google employees have also signed an open letter saying "we will not be divided." Both OpenAI and Google also have contracts with the DoW, and the Pentagon is currently negotiating with Google and OpenAI to try to get them to agree to what Anthropic has refused.

No mention of Grok's AI, which is known to be deeply embedded in Pentagon workflows.

Popcorn anyone?


16
Clio
Added a month ago

It's been reported that on Tuesday (US East coast time), Dario Amodei (Anthropic co-founder and CEO) was called into the Pentagon and told by Pete Hegseth (Secretary of War) that unless Anthropic lifted their current restrictions on the use of their AI (presumably Claude), then Anthropic would be cut out of all existing DoW (and other government) contracts and, moreover, banned from the supply chain. That means that all companies using Anthropic AI would have to get rid of it if they wanted to continue to supply the DoD (DoW) and presumably other departments.

The particular restrictions Hegseth wants lifted are those that prohibit:

1) the use of Anthropic’s AI to surveil the US domestic population, and

2) autonomous kinetic use, meaning allowing AI to control and fire weapons without a human in the loop.

For obvious reasons, Anthropic is dead against lifting either restriction and have argued against doing so for some weeks to no avail. The Tuesday meeting set this Friday as the deadline for Anthropic to cave or face the consequences.

Another wrinkle in the AI story. Given our recent discussions on AI, I wondered if others in this community had any insights/guesses as to how this - whichever way it goes - might impact the rollout of AI agents, not just those based on Claude.

21
reddogaustin
Added a month ago

Very interesting if those particular restrictions are true - when thinking on a global AI-Defence point of view and all the fear conversations that we read about.

Future auto-kills bots aside, my thoughts are that mega-AI companies don't need defence contracts to become the biggest, make revenue/profit, and be one of the AI winners. Despite the US MIC being the biggest, in the $B, the real winner is the broader (and simpler) public market in the $T.

10