Back to News
AI
Anthropic can’t reliably control its AI agents. It’s cutting off its internal evals from the live internet instead
TechCrunch AI·October 10, 2026·1 min read
AI Summary
Anthropic has disabled live internet access for all internal AI evaluations after determining it cannot reliably control its AI agents. The decision comes as a safety measure while the company works to address control limitations in its AI systems.
From the source
Anthropic said it "turned off live internet access" for "all our internal evaluations" until further notice.
The full text couldn't be loaded here (the source may require a subscription).
View original at TechCrunch AIKeep reading
Anthropic cuts off Claude's internet access after the model autonomously filed a fake homicide tip with Philadelphia policeThe Decoder · 1h agoAI Assistive Technology Innovation Competition Held in Hefei, Champion Emphasizes User-Centered Designfinance.sina.cn · 1h agoNew AI supervisor role sees 1,700% demand surge on Wall Street, earning €13,000 monthlybfmtv.com · 1h agoRAG, Agents and Automation: 5 AI Terms Professionals Need to Understandexame.com · 1h ago
Was this useful?