Lesson 86 · 2 min

Ask for sources, then open them

A source is only useful once you open it and see that it really says what the AI claims.

Ask a chatbot for sources and it will give some. The problem is that a model without web search can invent sources that look real: a believable title, a real-sounding journal, a link that goes nowhere.

Tools with web search or a deep research mode do better here, because they link to pages they actually found. Even then, the summary can stretch what the page says.

So do the 60 second check. Open the link. Find the sentence that supports the claim. Ask who wrote it and when. A government page, a university or a well-known newspaper beats an anonymous blog.

If you cannot find the claim in the source, treat it as unconfirmed. Tell the chatbot "This link does not say that" and see what it does. Often it corrects itself, and sometimes it invents again.

Do it now

The 60 second source check

  1. 1. Ask a factual question you care about, with the prompt below.
  2. 2. Open two of the sources it gives you.
  3. 3. For each, write yes or no: does the page really say it?
[YOUR QUESTION]
Give a short answer, then list your sources with links.
For each source, quote the sentence that supports your answer. If you are not certain a source exists, say so instead of listing it.

Quick check

An AI gives you a source with a link. What should you do before trusting the claim?

An AI gives you a source with a link. What should you do before trusting the claim?

More on this

Words from this lesson

Hallucination
When an AI states something false with confidence, such as a made-up fact, quote or source.
Deep research
A chatbot mode that searches many web pages over several minutes and writes a longer report with sources.
Grounding
Tying an AI's answer to real sources you supply or it looks up, so claims can be checked instead of invented.