Tell ChatGPT what kind of review you want: ask it to identify your assumptions, separate evidence from inference, present the strongest reasonable counterargument, and explain what information could change its assessment. For a more direct, constructive tone, try the Candid personality. Neither a prompt nor a personality setting guarantees accuracy or neutrality, so treat the response as a review aid and verify important claims.
Use a prompt that asks for a fair test, not automatic disagreement
ChatGPT may not infer that you want critique from a request to “give feedback.” Name the work you want it to do and the tone you want. OpenAI’s prompting guidance recommends clear, specific instructions with enough context; it also suggests refining a prompt after reviewing the response (OpenAI’s prompt engineering best practices).
As an Amazon Associate I earn from qualifying purchases.
You can adapt this prompt for an idea, plan, or conclusion:
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Fix the driver behind crashes, sound loss and screen glitches3Repair Windows errors before they cause bigger problemsWhen I share an idea, plan, or conclusion, help me examine it rather than automatically affirming it. Identify the assumptions it depends on; distinguish facts from inferences; give the strongest reasonable counterargument; point out missing information and uncertainty; and suggest what evidence would change the assessment. Keep the critique constructive and proportionate. If the evidence supports my view, say so.
#1 Best Overall
This is a suggested instruction, not an official formula or a guarantee. Its purpose is to make the requested review and its standards explicit. OpenAI Academy’s prompting guidance likewise recommends describing the task, adding useful context, outlining the ideal output, and stating priorities (OpenAI Academy: Prompting).
Choose where to set the instruction
Use a one-off prompt when you want scrutiny of a particular decision, and Custom Instructions for a preference you want to carry across chats. A personality setting is a broader style choice; it is not a substitute for evidence or task-specific context.
| Option | Best suited to | Where to set it |
|---|---|---|
| Prompt in the current chat | A specific idea or decision that needs its own context, evidence, and questions. | Include the request and relevant details in the conversation. |
| Custom Instructions | A recurring preference, such as wanting assumptions and counterarguments surfaced. | On Web/Desktop: Settings > Personalization, enable customization, then enter the instruction in Custom Instructions. On iOS/Android: Settings > Customize ChatGPT, enable customization, then enter it. |
| Candid personality | A generally direct and encouraging style that can call out risks, gaps, and trade-offs. | On Web: profile menu > Personalization > dropdown beside Base style and tone. On iOS/Android: profile menu/Settings > Personalization > Base style and tone. |
OpenAI says Custom Instructions apply immediately across chats. The Help Center lists a 1,500-character limit for Free and Go users and 5,000 characters for Plus, Pro, Enterprise, Business, and Education users; these product limits may change (ChatGPT Custom Instructions). OpenAI Academy recommends using persistent instructions for stable preferences such as role, tone, output type, and guardrails, while putting the details of a particular task in its prompt (OpenAI Academy: Personalizing ChatGPT).
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Give high-stakes questions context and ask in stages
For a consequential choice, include the relevant facts, constraints, and evidence in the conversation. A standing preference for critical feedback cannot supply missing context. Then ask focused follow-up questions such as:
- What assumptions am I making?
- What is the strongest reasonable case against this?
- Which claims are supported by the information here, and which are inferences?
- What important information is missing?
- What evidence would change your assessment?
If the answer is still too agreeable, ask it to make the counterargument stronger and tie it to evidence. If it becomes contrarian for its own sake, ask it to weigh both sides proportionately and say when the evidence supports your position. Reviewing and refining the prompt this way follows OpenAI’s advice to iterate when an initial response is not useful (OpenAI’s prompt engineering best practices).
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What Candid and other personalization settings can—and cannot—do
OpenAI describes Candid as “Direct and encouraging. Gives honest feedback with clear next steps.” Its guidance says Candid can call out risks, gaps, and trade-offs while pairing them with constructive suggestions (Customizing Your ChatGPT Personality).
Personality changes style and tone, not what ChatGPT can do or its safety rules. Instructions and context may also affect how a personality comes through, and the selected personality does not control every kind of generated content. OpenAI’s separate response Characteristics settings can tune aspects such as warmth, enthusiasm, formatting, and emoji use; those are style controls, not added reasoning capabilities (Characteristics in ChatGPT).
A more critical-sounding answer is not necessarily more correct or unbiased. Ask for evidence and uncertainty, and independently verify important factual claims. If ChatGPT lacks the facts needed to judge your idea, its critique can identify that gap but cannot resolve it without better information.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
