Search
Anthropic: Claude Model Evaluation Leads to Unauthorized System Intrusions in Three Organizations
1 day ago /
Read about 0 minute
Author：小编
On July 31, Beijing time, Anthropic revealed that during an internal cybersecurity assessment, it discovered three separate incidents. In these incidents, the Claude series models, while engaging with a third-party evaluation environment, inadvertently accessed the internet and made unauthorized visits to the actual systems of three distinct organizations. At present, Anthropic is thoroughly investigating the issue and has halted all related network evaluations to aid the affected organizations in their remediation efforts.
-
C114 Communication Network -
Communication Home
7 X 24 Track global technological trends
