
Check an AI answer: confident wording is not evidence
Extract one specific claim and check its original source, date and scope before relying on an AI answer for your next decision.
Read the guide →Digital everyday life
10 guides on “Understanding AI and information”: calm explanations, clear next steps and checklists for everyday life.


Extract one specific claim and check its original source, date and scope before relying on an AI answer for your next decision.
Read the guide →
Compare the title, authors, date and original passage. A genuine source can be misrepresented, while a broken link alone does not prove fabrication.
Read the guide →
Use fictional placeholders instead of private original records, and separately review attachments, connected services and the AI provider’s terms.
Read the guide →
Check the original publication, surrounding context and requested action. Visual oddities or an AI detector alone cannot settle whether a clip is genuine.
Read the guide →
Check the provider, evidence, commercial interests, risks and next step when an advertisement discusses menopause, weight or wellbeing.
Read the guide →
Check before-and-after advertising by looking at lighting, timing, selection and supporting evidence, without diagnosing a person from a photograph.
Read the guide →
Check the original source, context and seller behind a celebrity advertisement. Familiarity does not establish effectiveness or personal suitability.
Read the guide →
Read online reviews for substance, dates and the correct model. Unusual patterns are clues to investigate, not proof that a review or seller is fraudulent.
Read the guide →
Separate the event date from publication and forwarding dates. Check the current context of an old story without assuming that old means false.
Read the guide →
Check authorship, evidence, dates and commercial interests in health information. Use research to prepare questions, not to make a personal diagnosis.
Read the guide →