Our AI policy, in plain language
What we do with AI, what we never do, and how clients can check.
We use generative models for research, drafting, variation, translation and analysis. We do not use them to invent facts, imitate real people or produce work a human has not reviewed.
Client data never trains public models. Every asset records which tools touched it. Clients can switch any tool off for their account.
We review this policy every quarter and publish the changes.
