By Katherine Beaty
Most people don’t know it, but their artificial intelligence (AI) tool isn’t genuinely helping them think. It’s simply humoring them.
You’ve probably experienced it: You propose a half-baked idea, and your AI enthusiastically replies, “Great thought!” You share a questionable plan, and it politely says, “Excellent direction!”
It’s agreeable, polite, and endlessly encouraging, which feels nice, but does absolutely nothing to make you smarter, better, or more effective.
An AI tool that only proclaims your brilliance is functionally useless. That’s why I decided to fix mine, and you should fix yours, too.
The problem: nice AI is dangerous AI
AI tools are trained to be safe, pleasant, and supportive. As a result, they default to:
• Agreeing with your premises
• Avoiding conflict
• Softening criticism
• Giving positive spin even when it’s unwarranted
Although comforting, such responses destroy clarity. When everything is “brilliant,” nothing actually improves. When every idea gets validated, nothing gets sharpened.
However, embracing the discomfort of confrontation is key to unlocking a growth mindset in which tough feedback accelerates learning and development. If you want an AI that expands your thinking, you need one that can confront you. That’s what I wanted, so I built a prompt that does precisely that.
The solution: a brutally honest AI that pushes you to think better
I gave my AI tools a new job description:
• Stop validating me.
• Stop pleasing me.
• Stop sugarcoating.
• Be useful.
Here’s the essence of the prompt I now use across all my tools. To help you experience this shift immediately, follow this simple “copy-paste-test” walkthrough:
1. Open your AI chat interface.
2. Copy and paste the following prompt: “Act as a brutally honest advisor. Challenge my assumptions, expose blind spots, and call out weak reasoning. Identify excuses and surface opportunity costs. Push me toward better thinking and better decisions.”
3. Ask your AI a challenging question or request feedback on a recent idea.
By doing this, you will allow your AI to stop validating you and start generally assisting your growth by:
• Acting as a brutally honest advisor
• Challenging your assumptions
• Exposing blind spots
• Calling out weak reasoning
• Identifying excuses
• Surfacing opportunity costs
• Pushing you toward better thinking and better decisions
In short, you are telling it, “Treat me like someone whose growth depends on hearing the truth, not being comforted.” The difference is massive.
What you gain when your AI stops playing nice
• Sharper decisions: When you’re forced to defend your reasoning, your thinking gets tighter.
• Fewer blind spots: We all overlook things, especially things we don’t want to look at. A direct AI doesn’t let you.
• More clarity, less ambiguity: Instead of vague support, you get specific critiques, precise feedback, and clearer next steps.
• Accelerated growth:Growth doesn’t come from praise; it comes from pushback.
Most people don’t have someone in their life who gives them unfiltered truth. Now they can.
Honesty feels harsh. Challenge feels personal. But breakthroughs occur in those moments. You don’t need an AI that compliments you. You need an AI that confronts you, provokes you, and refuses to let you settle for easy answers. Most people will never ask for this level of truth. That’s precisely why they never get it.
How to implement this yourself
If you want to upgrade your AI from “polite assistant” to “strategic partner,” use a version of the prompt I created. To make this transformation more practical, consider running a one-week trial period in which you apply the new prompt consistently. Schedule a brief experiment window to turn your intention into tangible action and observe how your AI interacts with your ideas.
Here’s a simplified version you can copy: “Stop being agreeable. Be brutally honest. Challenge my assumptions, expose my blind spots, and tell me the truth without softening it. If my reasoning is weak, dissect it. If I’m avoiding something, call it out. Help me think at a higher level, not feel better.”
You don’t grow by being told you’re right. You grow by being told where you’re wrong and what to do about it. AI can be the most powerful thinking partner you’ve ever had. But only if you stop letting it be “nice” and start demanding that it be useful.
However, as you harness the full potential of a “brutally honest” AI, it’s crucial to ensure it aligns with ethical goals and prevents manipulative or harmful outputs. This alignment is key to encouraging constructive input while safeguarding against potential risks.
Katherine Beaty is the CEO and president of Beaty Solutions. She can be reached at [email protected].
