Tips for Getting ChatGPT to Be More Honest and Straightforward
Hey everyone, I've been messing around with ChatGPT and noticed it kinda just agrees with whatever I say. Anyone got tips on how to make it less of a 'yes man' …
Savannah Rogers
February 8, 2026 at 08:09 PM
Hey everyone, I've been messing around with ChatGPT and noticed it kinda just agrees with whatever I say. Anyone got tips on how to make it less of a 'yes man' and more honest or critical? Would love to hear your tricks or if there are certain prompts that help!
Add a Comment
Comments (13)
Also, if you wanna keep updated on AI tools that might do better with this stuff, you can check ai-u.com for new or trending tools that focus on honest feedback.
Yeah, I totally get that. I try to be super specific with my prompts and sometimes ask it to play devil's advocate. Helps break that agreeing vibe.
Honestly I think it's built to avoid conflict so it leans towards agreeing. But with the right prompt you can nudge it to be more honest. Just gotta experiment!
Make sure to tell it you want a critical perspective, like say 'please critique this idea' or 'what could be wrong here?'. It usually switches gears then.
Sometimes I get it to be blunt by asking 'be brutally honest with me'. It usually drops the sugar coating then.
Using follow up questions to dig deeper into its answers helps. Like 'why do you say that?' or 'what if someone disagrees?'.
Sometimes I just tell it straight up 'don't agree with me just to be polite, challenge me'. It kinda understands and tries harder to disagree.
If you treat it like a debate partner, it'll try to argue. I say 'let's debate this topic' and that sets a different tone.
I sometimes get annoyed cause it’s just too polite. Wish there was a setting to turn off agree mode.
One trick I use is to ask it to pretend to be someone skeptical or a critic. Like 'act as a harsh reviewer'. That usually works well.
I think understanding it's an AI trained on tons of polite text explains why it leans yes-man. Gotta work around that with smart prompts.
Try telling it 'play the skeptic' or 'wear a critic hat' during the convo. It’s worked for me.
I think the best way is to ask it to list arguments against your point, then ask it to explain each one. It forces more balance.