← Home

AI Notice

How Draftbench uses artificial intelligence

What is AI‑generated

Drafts, clause reviews, risk memos, revisions, contract summaries and key‑date extraction are generated by an AI model (Anthropic's Claude), guided by Draftbench's own instructions and checks. Every exported document is marked "AI‑assisted draft" with the date the legal information is current to ("law as of"), and carries a machine‑readable AI marker in its file properties. The AI model version may change as improvements are released.

What Draftbench checks automatically

Before a document can be marked final:

What AI can get wrong

AI can:

Research on leading legal AI tools shows errors remain common, so check every document before you rely on it. This applies especially to anything labelled "unverified", or where no recent legal updates are recorded for your governing law.

Human review

Your data

Your content is processed by our AI provider only to produce your output and is never used to train AI models. See the Privacy Policy, section 4.

No automated decisions about you

Draftbench produces drafts and analysis for you to use. It does not make decisions with legal or similarly significant effects about any person.

Tell us when it's wrong

Use "Report a problem" on any clause or finding. We review reports, and confirmed errors are used to improve the Service.

AI‑use statement

Need to tell a client or court that AI was used? Generate an AI‑use statement from any matter, listing which steps used AI and what was reviewed by a person.

Last updated 27 October 2026.