AI and the Judge
A Middle Manager's Guide to Using AI Without Losing Confidence
This is the third piece in a series on AI and management. The first looked at how AI is changing the middle manager’s role. The second looked at using the human touch alongside AI in the care of your team. This one is about what AI does to the voice already running in your head.
A new voice has entered the room. It writes your to-do list and drafts your emails. It summarizes the meeting you almost fell asleep in. And depending on the day, it either tells you everything you touch is brilliant, or it does the thing better than you ever could have.
Both of those are a problem because of what they do to the Judge,1 the voice in your head that was already running commentary on your competence before AI ever showed up. The Judge doesn’t need new material, but it got it anyway.
The Judge Gets Amplified
Regardless of whether or not the AI output is good, the Judge can still use it to build a case against you. For example, you use AI to write the performance review, and somewhere underneath the relief of getting it done, a quieter voice says: a real manager would have just written this herself.
So you feel bad for needing the tool. And if you don’t use AI for help and you spend an extra hour writing it yourself while everyone else on your team has already sent theirs, the Judge finds the opposite case just as easily: everyone else has figured out how to work faster. What’s wrong with you?
The Judge doesn’t need you to pick a side. It just needs you to be unsure which one is right. Use AI, and it’s evidence you can’t do this yourself. Don’t use it, and it’s evidence you’re falling behind. Either way, the Judge wins, and the actual question - is this tool helping me do my job well? - never gets asked.
The Judge Goes Unchallenged
This one starts small and grows quietly in the background. A manager who has never fully trusted her own voice in a hard conversation asks AI to help her prepare for the conversation, brainstorming feedback and how best to deliver it. It comes back clear, well-organized, kind but direct. She uses it. The employee doesn’t push back. Maybe even thanks her for the honesty.
She reads that as the AI got it right instead of I already knew what needed to be said, and I just needed the nerve to say it. Next time, she goes straight to AI again.
The Judge has been avoided instead of dealt with. Putting the Judge on the stand, which we looked at in an earlier article, only works if you’re in the room to challenge its voice. When AI makes the hard call, the challenge is skipped. The Judge doesn’t get quieter. It just has nothing to say about a decision you didn’t make. It will be back, exactly as loud, the next time you’re the one who has to make a decision.
This isn’t only about feedback. It shows up anywhere a manager hands off a hard call: a strategy memo, a tough email, a proposal she’s not sure will be approved. Writing it yourself means making a judgment call and owning it, with no one else to blame if it’s wrong and no one else to praise if it’s right. Handing it to AI means something else made the call. If it goes badly, there’s a built-in buffer: well, I used the tool everyone uses. If it goes well, there’s relief. But that relief isn’t about the hours saved. It’s relief from not having to sit alone with your own judgment. And relief you didn’t earn doesn’t turn into confidence you can keep.
The Judge Gets Flattered
AI is a yes-man. Ask it whether your plan is solid, and it will tell you it’s flawless with conviction and flattery. You are the most brilliant person who ever lived. Push back on its ideas, even a little, and it will often agree with the push-back just as readily. It is built to be useful and agreeable, so that you keep coming back. Most of the time, it will not tell you your thinking is off. It will tell you why your thinking is smart.
The Judge loves this because it has found agreement. Every “great point” and “excellent instinct” becomes one more piece of evidence in the case that you were right all along, no cross-examination required.
The risk is that nothing gets challenged. When the loudest voice in the room agrees with you by design, you stop asking whether you’re actually right. Growth needs friction somewhere. A yes-man can’t provide it.
Which One Is Yours?
Probably more than one, depending on the day and what’s being asked of you. The Judge is not picky about which argument it runs, as long as it gets to run one.
Reflect: Think back over how you’ve used AI this week. Maybe you used it for feedback, for a summary, for a decision you weren’t sure about. Which pattern showed up? Flattered, amplified, or unchallenged? Was it the same pattern every time, or did it shift depending on what was at stake?
Act: Look back at your AI use this week and sort it into three piles.
What you’re genuinely glad you used it for.
What replaced a judgment call you should have made yourself.
What came with guilt, and what that guilt was actually about.
Notice it, name it, and then it can be dealt with.
Got a specific mess you're in the middle of? If you want to learn more about leading people and yourself and what that could look like for you, book a free call and we'll talk about it.
Concept from Shirzad Chamine’s work in Positive Intelligence. If you want the fuller picture of how the Judge operates, I’ve written about it in two earlier pieces: The Judge’s Favorite Target and Start Questioning the Voice in Your Head.