B-AIM designs AI to support people while keeping accountability human.1. Purpose and benefit — define the user need, intended benefit and accountable owner.2. Evidence and honesty — cite reliable sources, label uncertainty, and separate scientific evidence from symbolic or traditional material.3. Human oversight — meaningful review, override and appeal remain available for consequential decisions.4. Fairness and accessibility — test for exclusion, bias and unequal impact across affected groups.5. Privacy and security — minimise data, protect access, and never publish credentials or private-client information.6. Transparency — explain AI involvement, limitations, provenance and material changes in plain language.7. Rights and consent — use only authorised content, identities and media; respect intellectual property.8. Safety and continuous review — monitor outcomes, record incidents, pause unsafe use and improve controls.No B-AIM concept should be presented as a guaranteed result. High-impact medical, financial, employment, legal or public claims require qualified human review.