How to Keep Your Judgment Sharp When You Work With AI

How to Keep Your Judgment Sharp When You Work With AI

A GPS once tried to drive me straight into a protest march in downtown Dallas. It kept telling me to turn onto roads that were clearly blocked to traffic. The machine had no idea the road was closed. It only knew what it knew. I was the one who could see the signs. 

Training and mastering sound judgment (checking what the tool knows and what you know, observing, and intuiting) is the most important skill for at work right now.

The cost of skipping discernment

A lawyer with three decades of experience used an AI tool to research a brief. It produced fabricated cases and citations. He even asked the tool to confirm the citations were real, and it told him they were. They were not. The court dismissed the case and the lawyers were fined. With 30 years behind him, he almost certainly had a few twinges that something was off. He ignored them. One lapse in discernment, and his credibility took the hit. That can happen to any of us, and it is happening to a lot of us.

The more you trust the tool, the less you check it

This might grab your attention.

Microsoft and Carnegie Mellon studied knowledge workers and found that the more confident people are in AI, the less they think for themselves!

But the good news is: the more confident they are in their own skill, the more they validate and double check.  (whew)

Harvard Business Review has made a related point: as tools get more powerful, human judgment becomes more important, not less. The machine raises the floor on speed. Your judgment is still the ceiling.

The danger is that the people who most need to question the answer, the ones least sure of the task, are the ones most likely to accept it and move on.

Don’t let that be you or your team.

• Build your self-trust & confidence
• Strengthen your discernment, start by just noticing and naming the flickers/intuition/feelings you sense
• Apply this simple tool, the Sign-Off℠

A 30 second gate before you sign off

You can't outsource judgment to AI, but you can build a habit that protects it. I use a quick check called the Sign-Off℠.

• Blind Spots: what did the tool assume that you didn't tell it, or what did it miss?
• Stakes: what happens if this is wrong, and how much does it matter?
• Signature: are you willing to put your name on it and accept the consequences?

If you stall on any of the three, pause and dig deeper. It works on AI output, and it works on a the data your team hands to you.

This week, pick one thing AI tells you, ideally the claim that would hurt most if it were wrong, and check it against another source. Then run the Sign-Off℠ on a real recommendation.

AI can hand you the answer fast. It cannot hand you the judgment to know if that answer is right for your people, your context, and your timing. That part is still yours. Protect it.

Watch this week's Mini but Mighty℠ Leadership Lesson on YouTube

If you are seeing managers across your team accept confident answers without checking, our workshops and courses build discernment and judgment at scale. Reach out at bossatthehelm.com/contact.

Lead yourself. Then lead others. Be a Boss.