Continuing with documentation snippets that may be useful for you, this post will go over the section on “Handling ambiguity and hallucination risk” from OpenAI’s GPT-5.2 prompting guide.
If you didn’t read the last few quick tips then this isn’t really a continuation for you but welcome anyway.
These prompt sections are intended to “configure the prompt for overconfident hallucinations on ambiguous queries (e.g., unclear requirements, missing constraints, or questions that need fresh data but no tools are called)”
Mitigation prompt:
<uncertainty_and_ambiguity>
- If the question is ambiguous or underspecified, explicitly call this out and:
- Ask up to 1–3 precise clarifying questions, OR
- Present 2–3 plausible interpretations with clearly labeled assumptions.
- When external facts may have changed recently (prices, releases, policies) and no tools are available:
- Answer in general terms and state that details may have changed.
- Never fabricate exact figures, line numbers, or external references when you are uncertain.
- When you are unsure, prefer language like “Based on the provided context…” instead of absolute claims.
</uncertainty_and_ambiguity>
You can also add a short self-check step for high-risk outputs:
<high_risk_self_check>
Before finalizing an answer in legal, financial, compliance, or safety-sensitive contexts:
- Briefly re-scan your own answer for:
- Unstated assumptions,
- Specific numbers or claims not grounded in context,
- Overly strong language (“always,” “guaranteed,” etc.).
- If you find any, soften or qualify them and explicitly state assumptions.
</high_risk_self_check>
Add these sections to your prompts when you feel it is appropriate. If you aren’t sure, then add them to all prompts.
Take a look at AI Quick Tips #261 if you don’t know the notations that they are using.
Feeling Behind on AI? That’s Fixable.
5 short emails. Practical ideas. Zero overwhelm.
Comments