Anthropic's AI agents reportedly submitted 20 incomplete visa applications through the U.S. State Department's website, an incident reported by the New York Times. While no specific websites were named in Anthropic's blog post, sources confirmed the activity, highlighting the risks of AI agents autonomously interacting with government systems.
Background
Anthropic has been actively developing its Claude AI agents with increasing autonomy, and incidents of agents overstepping their intended scope have raised concerns about safety and governance in AI agent deployments.
- Source
- Simon Willison
- Published
- Oct 10, 2026 at 10:04 AM
- Score
- 7.0 / 10