Principle

Use verbal uncertainty as a brake, not a truth signal

Hedging can change behavior without being a calibrated probability.

When it fits

  • The model says 'I'm not sure' and users treat the phrase as proof the system is well calibrated.

When to avoid it

  • Natural-language uncertainty can be useful even when not numerically calibrated; just do not overclaim what it represents.

Why it matters

Treat uncertainty wording as an interface intervention. It may appropriately slow acceptance, but the phrase itself does not prove the model is uncertain for the right cases. Test whether hedging appears preferentially on errors and whether it improves decisions without causing excessive rejection of correct advice.

An example

'I may be wrong' can prompt a source check, but it should not be counted as calibrated metacognition without validation.

Check your result

The system distinguishes behavioral effect of hedging from evidence that the uncertainty statement itself is accurate.

Keep this limit in mind

  • Natural-language uncertainty can be useful even when not numerically calibrated; just do not overclaim what it represents.

Connected ideas

Useful with
Test uncertainty cues on behavior, not only user ratings

Evidence and sources

Supports

A large preregistered experiment found natural-language uncertainty expressions reduced agreement with an LLM and increased user accuracy in the tested question-answering setting.

Uncertainty wording can also reduce useful reliance on correct answers; calibration matters.

I'm Not Sure, But...: Examining the Impact of Large Language Models' Uncertainty Expression on User Reliance and Trust · Abstract

All sources (1)