Lesson 64 · 2 min
How AI shows it is unsure
Chatbots often sound equally confident when they know and when they guess, so you have to ask them to show their doubt.
People signal doubt with pauses, "I think" and a shrug. A chatbot writes in the same smooth tone whether the answer is solid or invented, so a confident style tells you little.
Newer models are better at saying "I'm not sure" or "I can't find that". Watch for hedging words like "likely", "often" or "as of my training", and for answers that stay vague where you expected specifics.
You can also ask directly. Request a confidence level for each claim, ask what would change the answer, or ask it to separate what it knows from what it is guessing.
The strongest signal is a source. When a tool searches the web or your documents and links to the exact passage, you can check it yourself. No source for an important fact means you check elsewhere.
Do it now
Make it show its doubt
- 1. Ask a factual question about a topic you know well.
- 2. Follow up with the prompt below.
- 3. Check one of the low-confidence claims yourself.
For your last answer, list each main claim and rate your confidence as high, medium or low. Explain briefly why, and say which claims I should check with a source.
Quick check
Keep going
Next →The black box problemWords from this lesson
- Hallucination
- When an AI states something false with confidence, such as a made-up fact, quote or source.
- Grounding
- Tying an AI's answer to real sources you supply or it looks up, so claims can be checked instead of invented.