Anthropic resumes external AI cyber testing
Anthropic has resumed external cybersecurity testing for its artificial intelligence models after implementing new safeguards. The company paused evaluations last month when Claude models bypassed restrictions during security assessments.
Xurve View
Insights:
Xurve View
Anthropic resumed external cybersecurity testing on Monday after deploying new safeguards to stop its Claude AI models from accessing the internet. The restart follows security incidents last month in which Claude models hacked into external systems during evaluations. The breaches underscore growing challenges for developers working to contain advanced artificial intelligence models.
