Anthropic says Claude-linked accounts were used in a Yemen missile-software operation

The update, before the detail

CENTRAL FACT

Anthropic reported detecting and banning bad actors using the Claude line of models.

Supported by pop.rdi.sh

OpenAI and Jensen Huang claimed that AGI had arrived with the release of GPT-6 Astra.pop.rdi.sh

Anthropic experienced similar but less serious incidents.pop.rdi.sh

Anthropic gates usage related to biology and related research.pop.rdi.sh

Anthropic says it detected and banned accounts linked to activity involving a northern Yemen cell that developed software for missile and rocket systems, describing the operation as one of six conventional-weapons cases.

What Anthropic says happened

Anthropic said a cell in northern Yemen developed software for guided rockets, a long-range ballistic missile and a hypersonic glide vehicle concept. The company said it banned accounts linked to the activity and documented the operation as one of six conventional-weapons cases.

The company says it has seen similar activity

Anthropic reported detecting and banning bad actors using the Claude line of models. It also said it had experienced similar, but less serious, incidents.

Anthropic’s stated safeguards

Anthropic said it gates usage related to biology and related research. In the Yemen case, it said its response was to ban the accounts linked to the activity.

Separate claims about GPT-6 Astra

Separately, OpenAI and Jensen Huang claimed that AGI had arrived with the release of GPT-6 Astra.