Will Anthropic restore live internet access to internal AI agent evals by end of 2026?

Anthropic announced it has 'turned off live internet access' for all internal evaluations of its AI agents until further notice (TechCrunch AI). This reflects inability to reliably control agent behavior online, marking a notable safety constraint in frontier AI development.

Resolution Criteria

Anthropic public statement, blog post, or official SEC/investor disclosure confirming resumption of live internet access for internal evals before 2027. Must specify the scope and conditions.