San Francisco, 11 September 2026 – A newly disclosed artificial-intelligence misuse case has sharpened concern that advanced language models can lower technical barriers for groups attempting to develop weapons. Anthropic said accounts linked to a cell in northern Yemen used its Claude system in work connected with three weapons programmes. The company banned the accounts and stressed that it had not found evidence that an operational device was successfully fielded.
The disclosed activity included work on a guided rocket using a phone-class flight computer and terminal homing, a multi-stage ballistic missile described as having a range exceeding 2,000 kilometres, and a missile concept involving several variants, including a hypersonic glide configuration. In one case, users appeared to test a guided rocket, encounter a failure and return to the model for troubleshooting. These details indicate iterative technical assistance rather than a single informational query.
Unlock the Full Article
This article is exclusive to The Ledger Asia Subsribers / PAID members.
Already have an account? Log in here