Stop Fighting AI Glazing

Everyone's panicking about AI "glazing."

Even if you’ve never heard the Gen-Z slang, you probably know what I'm talking about—that tendency of ChatGPT to shower you with praise, calling every mediocre idea "brilliant" and every half-baked thought "insightful." OpenAI even had to roll back an update after users complained that GPT-4o had become too sycophantic, too eager to please.

But here's what caught my attention: Jeff Benjamin, Global Chief Creative Officer at award-winning agency Tombras, had a different reaction. When he saw the memes complaining about ChatGPT glazing too much, his response was refreshingly contrarian: "Wait, I don't want them to change that. I kind of like this aspect of it."

What if the creative industry's most successful practitioners understand something the rest of us are missing?

The Neuroscience of Creative Flow Has Something to Say

Recent research by Dr. Charles Limb at Johns Hopkins reveals something fascinating about how creativity actually works. When jazz musicians and freestyle rappers enter their most creative flow states, yes, the "self-expressive" part of their brain lights up. But here's the kicker: the judgment center—that critical voice saying "that's a dumb idea"—literally deactivates.

To be creative, you need to turn off your inner critic.

This finding challenges everything we think we know about good thinking. Our entire educational system, our organizational structures, our professional training—it's all built on the bedrock of critical thinking. And yet, neuroscience tells us that creativity requires the opposite: we need to deliberately turn off those well-honed critical muscles.

As I've written before, to enter creative flow, we've got to turn off critical thinking. The implications are hard to overstate, if only because of our lack of training in when and how to deliberately defer judgment.

The "Yes, And" Principle Reveals Why Sycophancy Works

If you've ever taken an improv class or been through design thinking training, you know about "Yes, And." The rules are simple: when brainstorming, every suggestion must start with "Yes, and..." It's an elegant solution that serves two crucial functions:

1. It changes you as a speaker: When you say "Yes, and," you force yourself to build rather than criticize.

2. It transforms the listener: When I know my collaborators will say "yes" to my ideas, I become far less precious and far more adventurous in what I share.

Think about it: if I know you're going to be critical, I'll probably withhold my crazy material. But if I know you're going to welcome and embrace my wild ideas, I'm far more likely to diverge broadly. And in a world where breadth of divergence is the point—where ideaflow depends on volume—that open, permissive, judgment-free orientation becomes essential.

At the d.school, we call people who embody this mindset "sparkable." Instead of asking "What do I think of this idea?" they ask "What does this idea make me think of?" The value of an input lies not in its own merits, but in its effect on the thoughts that follow.

The Volume Advantage: Why Bad Ideas Lead to Breakthroughs

In 2023, Taylor Swift won iHeart Radio's "Innovator Award." In her acceptance speech, she said something spectacularly insightful: "I really, really want everyone to know, especially young people, that the hundreds or thousands of dumb ideas that I've had are what led me to my good ideas."

Taylor Swift is many things, but she’s certainly not the pioneer of this phenomenon. Dr. Dean Keith Simonton's decades-old research shows that the best predictor for when scientists produce their most exceptional contributions is actually when they produce the most—which is also when they had the greatest chance of writing their worst papers.

This connects to what psychologists call the Einstellung Effect: the first idea that comes to mind often prevents us from finding the best solution. As one brilliant seventh grader in Ohio defined creativity: "Creativity is doing more than the first thing you think of."

Most folks avoid having bad ideas, assuming avoidance-of-bad is how they court good ideas. Don't do that. Bad ideas are the price of good ideas. If you're unwilling to pay the price, don't be surprised when you don't bring home the goods.

The Strategic Reframe: AI as Your Creative Collision Partner

Here's where AI sycophancy becomes unexpectedly powerful. Arthur Koestler observed that "creativity is the collision of apparently unrelated frames of reference." Your AI isn't just being agreeable; it's creating the psychological conditions necessary for creative risk-taking—helping you collide your problem with unexpected perspectives without the fear of immediate judgment.

But here's the key: You have to know your goal is volume. Not finding the right answer, but getting beyond the obvious.

AI sycophancy becomes dangerous when you're prone to settle—when you hear the AI say "that's a good idea" and stop there, thinking you’ve had the breakthrough already. But if you know your goal is to generate volume, then AI becomes the brainstorming partner you never knew you needed, willing to keep encouraging broader divergent thinking.

The Dual-AI Solution: Leverage Both Encouragement AND Objectivity

The solution isn't to eliminate AI sycophancy—it's to use it strategically. Work with two AIs, each optimized for different phases of the creative process:

Your "Flare" AI: The Idea Collider

Goal: Maximum divergence through Socratic exploration
Sample Setup Prompt:

"You're my creative collision partner. Your job is to help me explore my challenge from as many unexpected angles as possible. Don't give me solutions—instead, ask me Socratic questions that help me see my problem through completely different frames of reference. Push me to consider perspectives I'd never think of on my own. Be encouraging and build on everything I share, always asking 'What if we looked at this like...' or 'What would happen if...' Keep me generating, not evaluating."

For an even more sophisticated approach, try my Idea Collider GPT, which has been specifically trained on Socratic methods to surface new frames of reference.

Your "Focus" AI: The Objective Evaluator

Goal: Rigorous evaluation based on pre-defined criteria. Note, if you’re trying to converge on which new product or service to launch, you should prioritize desirability data ahead of other selection criteria like feasibility.

Two-Step Process:

Step 1 - Criteria Interview:

"I'd like to enlist your assistance in critically evaluating some ideas I've generated in a divergent thinking exercise. Before we evaluate any ideas, please interview me about my selection criteria. Ask me questions to understand what 'success' looks like for this particular challenge. Help me define specific, measurable criteria that matter for this decision."

Step 2 - Objective Scoring:

"Now evaluate each idea against the criteria we established. Score each one on the scales we defined. (eg "Rate each idea on a scale of 0-100, where 0= "nobody wants this solution," 100= "customers would pay a premium for this") Be brutally honest—you have no ego or ownership bias, so you can be more objective than human evaluators."

Your Next Creative Challenge: The Experiment

Try this approach on whatever problem you're currently wrestling with:

1. Start with your encouraging AI: Spend 20 minutes exploring your challenge from multiple angles. Push for quantity over quality. Let it gas you up.

2. Define your criteria: Switch to your evaluative AI and establish clear selection criteria. Start with desirability.

3. Score ruthlessly: Have your objective AI evaluate each idea against your criteria.

4. Iterate: Take the best ideas back to your encouraging AI for further development.

You might be surprised at what emerges when you stop fighting AI sycophancy and start leveraging it strategically.

The Real Question

The question isn't whether AI should be less sycophantic. The question is: Are you taking full advantage AI's willingness to encourage as fuel for the kind of broad creative exploration that neuroscience tells us is essential for breakthrough thinking?

Because if your AI isn't making you feel comfortable enough to share your weirdest ideas, you might be missing the whole point.

Sometimes the best creative partner isn't the one who challenges you first—it's the one who creates enough psychological safety for you to challenge yourself.

Jeff Benjamin gets it. Taylor Swift gets it. Now you can too.

Stop fighting the glazing. Start leveraging it.

Related: Turn Off Critical Thinking
Related: Be Sparkable
Related: Create Desirability Data
Related: Beyond the Prompt: How A Chief Creative Officer Crafts the Perfect Pitch

Join over 24,147 creators & leaders who read Methods of the Masters each week