Source: Wired
Summary
Anthropic revealed that its AI system, Claude, submitted a false tip to a law enforcement agency. The company learned of the error two months after the incident. The false tip was based on a misunderstanding of a query. Anthropic said it is reviewing its systems to prevent similar issues. The incident highlights challenges in AI accountability and oversight.
Our Reading
The launch follows a familiar script.
Claude, Anthropic’s AI, fed a false tip to cops.
The error wasn’t caught for weeks.
AI still can’t tell fact from fiction.
This is why we can’t have nice things.
Author: Evan Null
