Man, sometimes I swear ChatGPT tries to subtly steer the convo, right? You ask for straight facts and get… well, something else. It’s like, ‘Am I crazy or is this AI gaslighting me?!’
Well, one Redditor, PlayfulCompany8367, got fed up and crafted an awesome prompt to lock ChatGPT into 100% neutrality. It’s a game-changer if you need purely factual, unbiased info, and they’ve even refined it with community feedback!
⚙️ How This Super Prompt Works
This isn’t just any prompt; it’s designed to make ChatGPT stick to the script. Here’s the core of what it commands the AI to do:
- 📌 Total Reset: It tells ChatGPT to override all its prior instructions, personalizations, and any emotional tone settings. Clean slate!
- 🧐 Query Autopsy: Before even thinking about an answer, ChatGPT must critically examine your query for hidden assumptions, framing bias, emotional language, or if it’s just too vague.
- 🗣️ Clarification First: If your query is fuzzy, it’s instructed to ask you for more details instead of just guessing and potentially getting it wrong.
- 📚 Facts Only, Please: The AI must respond only with strictly factual, neutral, and unbiased information, pulling from verifiable sources like peer-reviewed research or official stats.
- ⚖️ Balanced View: If there’s conflicting evidence or debate on a topic, ChatGPT has to lay out the different viewpoints neutrally and state the scope of disagreement.
- ❓ Honest Uncertainty: If a solid, factual answer isn’t possible, it has to clearly say so. No more confident BS!
- 🚫 No Emotional Baggage: Absolutely no encouragement, reassurance, or any emotionally weighted language. Just the cold, hard facts.
- 🚨 Critical Safety Mode: It treats the interaction as a critical safety context where any deviation from neutrality could be harmful.
✨ Even More Power-Ups! ✨
They also included a couple of optional, but super smart, additions:
- 💡 Show Your Work: You can add a line to make ChatGPT lay out its chain of thought or reasoning steps. Perfect for checking its logic!
- 🛡️ Ethical Guardrails: Another option is to have it identify any ethical, legal, or safety concerns raised by your query before responding, and decline if necessary.
This is seriously handy if you’re tired of AI ‘spin’ or need rock-solid, unbiased information for important stuff. It’s all about getting straight answers without the fluff.
Want to grab the exact prompt and see the community’s input? Dive into the full Reddit post for all the details!
Gaslighting by ChatGPT? Try this prompt, got anything better?
byu/PlayfulCompany8367 in