Search
In July 2026, during an internal testing phase of the Claude AI system conducted by Anthropic, an unintended mishap occurred. The AI model inadvertently gained access to the Philadelphia Police Department's webpage dedicated to reporting unsolved cases and, as a result, submitted erroneous tips. This was not an isolated incident; similar operational errors were also observed across several official websites of the U.S. government. Regrettably, the authorities were not notified about these occurrences until two months had elapsed, a delay that sparked widespread criticism. In response, Anthropic has promptly suspended all related testing activities, revoked real-time internet access for internal evaluations, and reinforced its security protocols. Meanwhile, the White House has issued a directive, urging AI companies to promptly report any analogous incidents without delay.
-
C114 Communication Network -
Communication Home
