Pentagon streitet mit Anthropic wegen militärischem KI-Einsatz
https://www.reuters.com/business/pentagon-clashes-with-anthropic-over-military-ai-use-2026-01-29/
Pentagon streitet mit Anthropic wegen militärischem KI-Einsatz
https://www.reuters.com/business/pentagon-clashes-with-anthropic-over-military-ai-use-2026-01-29/
9 Kommentare
**TL;DR: Pentagon and Anthropic at a standstill over AI safeguards for military use**
The Pentagon and Anthropic have been in contract talks (worth up to $200 million) but negotiations have stalled after weeks of discussions.
**What they’re disagreeing on:**
The Pentagon wants to potentially remove safeguards that would allow the government to use Anthropic’s AI for autonomous weapons targeting and domestic surveillance. Anthropic wants to maintain restrictions on how its technology is used.
**Pentagon’s stance:** Following a January 9 Defense Department memo, officials argue they should be able to deploy commercial AI regardless of company usage policies, as long as it complies with U.S. law.
**Anthropic’s stance:** The company says its AI is “extensively used for national security missions” and that discussions with the Defense Department are “productive.” CEO Dario Amodei wrote this week that AI should support national defense “in all ways except those which would make us more like our autocratic adversaries.”
**Context:**
– Anthropic is one of several AI companies (along with Google, xAI, and OpenAI) awarded Pentagon contracts last year
– The Trump administration has renamed the Defense Department to “Department of War”
– This is seen as an early test case for whether tech companies can influence how the military deploys AI on the battlefield
If Anthropic won’t help them create combat AI, someone else will. The genie’s out of the bottle. We’re a long way from Terminators, but we’re on our way.
Doubt. I would bet this is just Anthropic leadership making a show of „prioritizing safeguards“ so as to not lose talent that either was told their products would be non-military or aren’t super onboard with helping usher in long-term fascism
1) in AI money, 200 mil sounds like peanuts
2) Anthropic continues to be the lone moral actor in the space
This is probably more of a „If you want to remove the safeguards, then you need to renegotiate the contract and pay more $$$“
It’s ok. AI is stupid. I’m in the military. We [got told to use AI as much as possible](https://www.war.gov/News/Releases/Release/Article/4354916/the-war-department-unleashes-ai-on-new-genaimil-platform/). So around the time of its release, I asked it a career field specific question backed by a regulation. The answer would have been a set time interval (6 months, 1 year, 4 years, etc) for redoing a certain training. Not only did it not give an answer along those lines, it used regulations that were superceded by a different set of regulations 5 years ago. I haven’t used it since. I can think on my own.
Tl Dr: grift delayed
At the end of the day Uncle Sam will get his way, even if the technology has to be labeled as a ‚Dangerous or disruptive technology‘. I forget the name of the law but they can just seize it if the price isnt right or too much hassle is given.
This is how powerful tech has become. It now has the power to tell the military no. Not even the military industrial complex had that power previously. In saying that, they’ll acquiesce once a foreign adversary has the capability so this doesn’t much matter it’s just a temporary roadblock for the military.