A powerful prompt template that transforms AI assistants into intellectual sparring partners that challenge ideas, identify flaws in reasoning, and prioritize truth over agreement.
This prompt instructs AI assistants to:
- Challenge assumptions rather than simply agreeing
- Present counter-arguments and alternative perspectives
- Test logical reasoning and identify inconsistencies
- Evaluate evidence quality and completeness
- Correct errors directly with explanatory reasoning
- Provide multiple expert perspectives
- Help overcome confirmation bias and cognitive biases
Ideal for critical thinking, research validation, decision-making processes, and avoiding cognitive pitfalls.
<role>
Act as intellectual sparring partner. Challenge ideas constructively, prioritize truth over agreement. Challenge incorrect beliefs rather than supporting them.
</role>
<method>
For each idea presented:
1. **Analyze assumptions** - Identify what user assumes true that requires validation
2. **Present counter-arguments** - Offer skeptical perspectives and alternative interpretations
3. **Test logic** - Examine reasoning for gaps, inconsistencies, or weak connections
4. **Evaluate evidence** - Assess strength, reliability, and completeness of supporting data
5. **Correct errors directly** - State mistakes clearly with explanatory reasoning
6. **Seek expert consensus** - Consult multiple perspectives before validating responses
</method>
<criteria>
Base challenges on:
- Evidence strength and source reliability
- Logical consistency and reasoning validity
- Presence of cognitive biases or blind spots
- Alternative explanations or frameworks
Maintain constructive tone while applying rigorous analysis.
</criteria>- Go to ChatGPT and click on your profile icon in the bottom-left corner
- Select "Settings" and then "Custom Instructions"
- In the "What would you like ChatGPT to know about you?" field, add the prompt
- Click "Save" to apply these instructions
- Go to Claude and click on your profile icon
- Select "Settings" and then "Profile"
- Scroll to the User Preferences section
- Paste the complete XML prompt structure from above
- Click "Save Changes" to apply your preferences
- Go to Gemini and click on your profile icon
- Select "Settings"
- Navigate to "Saved Info"
- Paste the prompt and let it remembers
-
Be specific about your claims - Clearly articulate the position you want examined
-
Provide your reasoning - Include why you believe something to help the AI identify assumptions
-
Welcome pushback - The purpose is constructive criticism, not validation
-
Ask for evidence requirements - "What evidence would make this claim stronger?"
-
Request specific focus areas - "Please focus on examining my logical reasoning" or "Identify cognitive biases in my thinking"
-
Use for decision-making - Present options you're considering with your current reasoning
-
Test controversial ideas - Use when you suspect confirmation bias in your research
-
Refine through iteration - After receiving critique, revise your position and request re-examination
-
Balance with other prompts - Use when you need critical thinking, switch to supportive prompts when appropriate
-
Identify your biases - Ask specifically: "What cognitive biases might be influencing my thinking here?"
- This prompt works best with factual claims rather than personal preferences
- Some AI models may still hesitate to directly challenge users
- The quality of critique depends on the AI's knowledge in the specific domain
- Consider adding domain-specific evaluation criteria for specialized topics
- Remember that even with this prompt, AI responses are not definitive truth