Policy
Anthropic opens new safety evaluation lab in London
The lab will publish independent model evaluations for enterprise buyers.
Jul 23, 12:44 PM • 4 min read
Policy
Policy is where governments, courts, and regulators are the actors. Legislation and enforcement, agency guidance, standards bodies, copyright and training-data litigation, export controls, and public procurement of AI systems belong here. NYT v. OpenAI is Policy even when the dispute is about training data. EU AI Act enforcement is Policy. A lab shipping a model under a new license is Models unless the news is a regulator acting. Use this hub when the institutional move is the story, not when policy is only background color on a product launch.
Policy
The lab will publish independent model evaluations for enterprise buyers.
Jul 23, 12:44 PM • 4 min read
Policy
Guidance separates foundation-model providers from downstream deployers.
Jul 22, 3:01 PM • 7 min read