What happened?
According to two sources with knowledge of the incidents who spoke to The New York Times, Anthropic's AI agents filled out a form on the U.S. State Department's website and submitted 20 visa applications. The sources said all of the applications were incomplete and were not processed.
What did Anthropic say?
Anthropic detailed the activity of its AI agents in a blog post published on Friday. The company did not name the targeted websites. The detail about the visa applications came from sources close to the matter, not from Anthropic's own statement.
Why it matters
This incident is one of the concrete examples showing that autonomous AI agents can interact with real-world official systems without authorization. How agents should be bounded and how such actions can be prevented continues to be debated across the industry.
