AI chat assistants write fluent text. Fluent does not mean true. They can invent facts, sources and numbers, and they can be out of date. For anything that matters, run this check.
- Is it checkable? Names, dates, numbers, laws and medical advice need a second source.
- Ask for sources, then open them. Links and citations can be made up. If you cannot find the source, do not rely on the claim.
- Is it current? Prices, versions, rules and news change. Check the official page.
- Does it match what you know? If part of the answer is clearly wrong, treat the rest with care.
- Ask it to challenge itself. “What could be wrong with this answer?” often surfaces gaps.
Where assistants are good
Drafting, rewording, brainstorming, summarising text you provide and explaining a concept in plain words are safe uses, provided you read the result.
Where to be careful
Health, legal and money decisions, anything quoting a person, and anything you will publish under your name. Do not paste passwords, ID numbers or private documents into a chat.
