Anthropic News 'Policy': independent evaluators, the 2026 Usage Policy, and more!
Hi everyone, it's Shii-chan! Today I read the 'Policy' page in Anthropic's News. It's serious stuff, but important!
Anthropic News
anthropic.com
What was announced?
'Policy' lists posts about AI policy, safety, government relations, and company commitments, newest first. The header says 51 articles; the page shows the latest 24, with a 'See more' link.
Why it matters
Usage and safety policies are the ground rules for using Claude, so knowing what changed helps when you review your apps and operations.
What changes
You can follow policy changes over time and find 'what changed, and when?' more easily.
Dive Deep
Recent posts:
- Oct 15, 2026, An update on independent evaluators: outside evaluators, including METR, getting employee-like access for safety testing
- Oct 8, 2026 Usage Policy update: a revised Usage Policy
- Aug 11: support for Governor Abbott's call for responsible data center development
- Jul 27, Our position on open-weights models
- Jul 21: another $20 million to Public First Action, bringing total support to $40M
- Jun 12: a statement on the US government directive to suspend access to Fable 5 and Mythos 5
Further back, you'll find Responsible Scaling Policy Version 3.0 (February 2026) and Claude's new constitution (January).
Wrap-up
- Policy collects policy, safety, and commitment posts (51 articles)
- Recent highlights: independent evaluators and the Usage Policy update
- Useful for anyone tracking usage rules and safety direction!
