Anthropic embeds engineers in NSA, faces Pentagon fight
SkimNews Take
Musk's petition may be a legal maneuver to shed regulatory obligations by arguing a corporate entity has ceased to exist, rather than a genuine reflection of operational mergers.
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic embedded around six engineers inside the NSA to assist deploying its Mythos AI model for offensive cyber operations.
- Anthropic is engaged in a legal dispute with the Pentagon over the deployment of its Claude AI model.
- DoD denied Anthropic's request to reverse a blacklisting of its AI technology.
- Anthropic reports that over 80% of new production code in its codebase is generated by its Claude model.
- Anthropic publicly urged a global pause on AI development, citing risks from self‑improving models.
- Sam Biddle highlighted the inconsistency of Anthropic condemning Chinese AI hacking while now aiding the NSA's cyber operations.
Why it matters: The NSA gains a powerful AI‑driven cyber‑offensive tool, while Anthropic’s request to reverse its blacklisting was rebuffed by the Pentagon, limiting the company’s deployment options. The episode also fuels broader industry concerns that self‑improving AI could outpace oversight, prompting calls for a pause.
