AI assistants can produce wrong facts, outdated information and invented sources, all in the same confident tone as correct answers. A short fact-checking routine protects you.
1. Decide how much checking it needs
Brainstorming ideas? Little checking needed. Health, money, legal, news or anything you'll publish? Check carefully.
2. Ask for sources
For each factual claim in your answer, give the source and a link,
and say how confident you are. Mark anything you're unsure about.
Then open the links. Make sure they exist and actually say what the AI claims.
3. Check numbers separately
Statistics, dates, prices and calculations are common failure points. Verify them with official or primary sources, and redo important math.
4. Cross-check with a second source
Search for the key claim yourself, or ask a different AI tool and compare. Disagreement is a signal to dig deeper.
5. Watch for red flags
- Very specific details without sources
- Quotes attributed to real people
- Recent events (the model may be out of date)
- Answers that perfectly confirm what you hoped
Use search-connected tools for current events
Tools that search the web and cite sources are better for recent news, but still check the citations.
Read more: How to stop AI from making things up.