AI use policy

This policy sets the standard for Britain Next. It explains what readers can expect from us and how we protect public trust.

• Where AI is used: Ask Britain composes an answer to your question only from Britain Next's verified records and lists the records it used. Write to your MP drafts a letter from the answers you give. Our review officer checks challenges to published claims against their sources, and an analysis step sorts new official notices by topic and risk before a rules gate decides what publishes. Nothing else you read on Britain Next is generated by AI.
• Provider: these features run on Anthropic's Claude models under Anthropic's commercial API terms. Britain Next holds no account about you with Anthropic and does not store your questions, your answers or the drafts.
• Limits: the model is instructed to use only the supplied records, to say when Britain Next does not hold something, to report a vote in Parliament's own term (Aye or No), and never to guess who you meant. It can still make mistakes. Every answer carries links so you can check it against the record.
• People decide: no AI publishes a claim about a person's conduct or finances, applies a correction of that kind, or sends a letter. Those wait for a person.
• Challenge it: if an AI answer is wrong, email hello@britainnext.co.uk quoting the question and the answer. We log it and correct in the open under the corrections policy.