← Board·Strategy·importance 3/5 (editorial)
Report: Anthropic can’t reliably control its AI agents. It’s cutting off its internal evals from the live internet instead
- Current status
- Reported
- Companies
- Anthropic
- Happened
- Oct 9, 5:18 PM PT
- Last updated
- Oct 9, 5:20 PM PT (revision 1)
- Evidence
- 1 report · 1 independent chain (adjudicated)
What happened
TechCrunch reported: "Anthropic can’t reliably control its AI agents. It’s cutting off its internal evals from the live internet instead".
Why it matters
A change in how Anthropic operates affects its customers, users and competitors.
Uncertainty
Reported by TechCrunch; not independently confirmed.
Evidence and provenance
- First reportchain: techcrunchAnthropic can’t reliably control its AI agents. It’s cutting off its internal evals from the live internet insteadTechCrunch · Tim Fernholz · published Oct 9, 5:18 PM PT · techcrunch.com · rev 1