Permitted assistance
AI may help structure drafts, identify repeated language, propose clear explanations, generate code and support testing. It is not treated as an authority for legal, policy or factual claims.
Controls
- Check time-sensitive claims against current primary or authoritative sources.
- Do not fabricate people, qualifications, reviews, legal entities, identifiers, funding or access.
- Do not publish confidential information in prompts or outputs.
- Keep a human accountable for publication, correction and removal decisions.
- Disclose material AI assistance in the editorial policy.
Limitations
AI can produce plausible errors, miss context and reproduce bias. Human checking reduces risk but does not guarantee correctness.