This fall, a student in your district or on your campus will get caught letting ChatGPT write an essay. There will be a meeting, a policy citation, maybe a zero. That same week, someone in your cabinet will let ChatGPT write a board memo. That one gets a thank-you. We built entire academic integrity policies on one simple idea: if you didn't do the thinking, you don't own the work. Then we quietly exempted ourselves. Try this before your next cabinet meeting. Pull up the last recommendation someone on your team sent you: the budget reduction scenarios, the enrollment projection summary, the draft revision to the attendance policy. Ask the person who wrote it to walk you through the reasoning behind their second option, from memory, with the document closed. If they hesitate, you've found something your AI rollout plan never measured. Your team is getting faster at producing work it can no longer explain. All fall, I've been asking superintendents and college presidents the same question as they plan for 2027: what are your people carrying into this year that no training or workshop seems to touch? The first answer keeps coming back as one line. Convince me this is still worthy work. The second answer, more often than I expected, is AI. Leaders can feel the work speeding up and their own grip on it loosening. In a 2025 MIT Media Lab study , researchers put EEG caps on 54 people and had them write essays. The group using ChatGPT showed the weakest brain connectivity of anyone in the study. Minutes after finishing, most of them couldn't accurately quote a single sentence from the essay they had just turned in. The researchers named the pattern cognitive debt: you save effort today and pay it back later in thinking you can no longer do. It's a small study, and the researchers themselves asked people to stop calling it proof that AI makes anyone dumb. Its message for leaders is simpler. Every prompt that does your thinking for you is a small withdrawal from the one asset your board actually hired. Every cabinet I know is using AI more this fall. Very few are getting smarter because of it. The Irony Nobody at the Board Table Has Named Both halves of our profession are under the same public question right now: can the people we serve still think? In Gallup's latest annual education poll , Americans rated K-12 schools worst on critical thinking out of the six areas measured. Twenty percent called the job schools do there excellent or good. Forty-seven percent called it poor. Presidents hear their own version from employers, legislators, and parents writing tuition checks, all asking whether a degree still proves a graduate can reason through a hard problem. Now picture the same week inside a typical cabinet. The strategic plan update was drafted by a chatbot and lightly edited for tone. The enrollment memo came back from AI, got a quick read, and went to the board. The accreditation narrative or the state report got the same treatment. Someone asked the tool which of three budget scenarios to recommend, and the recommendation it gave is now the cabinet's position. We are asking our schools and colleges to teach students to think while the adults at the top quietly hand their own thinking to a tool that has no stake in whether the institution succeeds. (Nobody at the table would say it out loud. It's awkward to admit the drafts got better while your ability to defend them got worse.) Lee Lambert , who leads the Foothill-De Anza Community College District , told me on Season 1 of The PROOF that he thinks of AI as his tenth Supreme Court justice. He won't panic about it, and he won't ignore it. That's the posture I want for every cabinet, K-12 or higher ed. Give the tool a seat at the table. Keep the gavel. A 2025 study from Microsoft Research and Carnegie Mellon , based on a survey of 319 knowledge workers, found the same pattern at work. The more people trusted generative AI to handle a task, the less critical thinking they reported putting into it. Confidence in the tool went up. Effort in the human went down. Here's the part that should reach you personally. When the tool thinks for an individual, the damage stays in one head. When it thinks for a cabinet, the damage spreads. Five people defer to the same draft, nobody owns the reasoning, and the meeting turns into a line edit of a document nobody built. I've spent years in the chair and have since sat in many presidents' cabinet rooms, and the lesson holds in both: a cabinet that can't explain its own decision will eventually be unable to defend it. (This is the gap THE TEAM INSTITUTE was built to close. Every vendor will train your people on the tool. We help cabinets build the collective judgment the tool is supposed to serve, so AI multiplies what your team already thinks. More on that in a moment.) TQ = IQ × EQ × PQ. AI can push the IQ side of your output up fast. It can't supply the EQ your team needs to disagree well, or the PQ, the perceptual accuracy, that tells you when a polished draft is wrong for the people in front of you. Let those two drift toward zero and every gain on the IQ side gets multiplied by almost nothing. The Human Intelligence Protocol Call this the Human Intelligence Protocol. Five practices that turn AI into a sparring partner. Each works for one leader and gets stronger when a whole cabinet runs it. Protect the Genius Hours. Sort all of your team's work into two zones. The compression zone is anything large, messy, or repetitive that needs to become smaller and clearer: meeting notes, update memos, survey comments, three conflicting reports on the same program. AI is excellent here. The contribution zone is where your judgment comes in: deciding what matters, reading the room, noticing the issue nobody has said out loud, and knowing that an answer can be technically correct and still wrong for this community. The mistake is handing AI the contribution zone because it can produce something that looks finished. A better prompt before your next cabinet meeting: "Organize these notes and updates into key decisions, risks, unresolved questions, and missing information. Do not make a recommendation." The tool organizes the landscape. The room still chooses the direction. Climb the Brief Ladder. Most leaders use AI like a drive-through window: "Write my board update." Vague request, average answer, polished and forgettable. The ladder has five rungs. The request says what you want. Context says who it's for and where it will land. The outcome says what it must accomplish. Constraints set the boundaries, like length, tone, and what to avoid. The standard shows what good looks like: an example, a rubric, your own best past work. The brief ladder is also the skill of delegating to a vice president, a provost, or an assistant superintendent. If you can't brief the tool clearly, you probably haven't briefed your team clearly either. (The quality of your prompt is a pretty honest mirror of the quality of your thinking.) Seat a Board of Minds. A good board asks the hard questions. Left alone, AI tends to tell you your idea is wonderful. So give it other seats at the table. The stakeholder board: "Name the five groups most affected by this decision. For each, tell me what they want, what they fear, and what they'll resist." (Parents, teachers, and the union in one chair. Faculty senate, students, and donors in the other.) The time board: "What does this decision look like after one week, one year, and five years?" The opposition board, which matters most: "I believe this is the right strategy. Build the strongest possible case that I'm wrong. Name my false assumptions and the conditions that would make this plan fail." This one has a team payoff most leaders miss. In a cabinet, dissent carries a social cost, so it often goes unspoken. When the tool plays the dissenter, nobody has to spend political capital to raise the hard objection, and the room can argue hard about the idea while the relationships stay intact. One rule stays firm: AI states wrong facts with total confidence. When a decision rests on facts, ask for sources and verify the important ones yourself. Build a Pressure Lab. Pilots train in simulators. Emergency teams run drills before the crisis. Most leaders walk into the budget hearing, the tense board work session, the faculty senate meeting, or the hard personnel conversation having rehearsed nothing but their talking points. Use AI as the simulator. "Act as a skeptical board member reviewing our proposed program cut. Ask one question at a time. Push back on vague answers and raise the difficulty as I respond. Don't coach me until we're done." Afterward: "Evaluate my answers for clarity, evidence, and conciseness. Show me where I sounded defensive." Then run it again. The same works for learning. Ask for friction on purpose: "If I get stuck, give me the smallest useful clue." Letting the tool remove every difficulty is like having a friend lift your weights at the gym. Use the Beginner Advantage. The higher leaders climb, the more expensive it feels to say "I don't understand." So they nod through the fund balance discussion, the CIO's pitch about AI agents, the retention dashboard, or state assessment data nobody has explained since the last accreditation visit. Then they search for the answer later, or never. AI gives you a private place to go back to basics. "Explain this with no assumed background, in plain language, with one example from a school district or a college." "What are the five terms I need before anything else makes sense?" "What do beginners usually get wrong?" Then run the teach-back test. Close the window, explain the idea out loud in your own words, and ask the tool what you got wrong. Tell it to help you repair your explanation without rewriting it. When the boss does this openly and says so in cabinet, basic questions become normal for everyone. What This Looks Like in a Real Team This is a composite, drawn from patterns across our 987 teams. I've watched it play out in district offices and in presidents' suites, and the only real difference is the vocabulary. A cabinet had three weeks to bring a painful reduction plan to its board. (In the district version, two programs were on the table. In the campus version, two academic programs were up for sunset review, and the faculty senate was already asking questions.) The first draft came from AI in an afternoon. It was clean, confident, and organized. In the first internal review, the board chair's likeliest question ("Why this program and not the other one?") got four different answers from four different senior leaders. So they changed the process. AI compressed six months of budget data into a two-page briefing with the open questions labeled. The cabinet argued out its own recommendation before any tool saw the problem. Then they seated the opposition board, asked for the strongest case against their plan, and found two assumptions they couldn't support. Finally, the top leader and the CFO each spent twenty minutes in a pressure lab with an AI playing their toughest board member. On the night of the vote, the board chair asked the hardest question, and every cabinet member who spoke gave the same answer in their own words. The next morning, the team could explain the plan to the people it affected. Same tool. Same people. What changed was the order: think first, then let the machine argue with you. Bad Season or Wrong Seat? Some of what leaders are feeling this fall is the AI speed-up. Some of it is older than any chatbot. When the work gets heavy, most of us carry one question we rarely say out loud: is this a bad season, or did I take the wrong seat? Season 1 guests of The PROOF answered that question on camera. Rick Cobb waited out a bad year, and the next one became one of his best. Amy Diaz, EdD saw that the problem was the shape of her role and stepped sideways into work that fit. Chris Lindholm runs the district with the longest school bus ride in America. In Everton, Missouri, 91% of Will Darter, Ed.D students live in poverty, and his campus tripled in size while enrollment dropped. On Wednesday, October 7 at 10:30 AM CT, Season 1 guests join our P2P Roundtable to talk it through live. One hour, peers only, no presentations, 30 seats, open to anyone leading systems and people in education. The question sounds the same in both chairs. You'll leave inspired, challenged, and equipped to lead as a BUILDER (not a climber, dreamer, or coaster). Save your seat: higherperformancegroup.com/p2p-page Bring your closed-document test result with you. It's a good place to start the conversation. Monday Morning Here's what to do Monday morning (assuming you're not already mid-crisis, in which case bookmark this and do it Tuesday). Step 1: Run the Zone Audit (20 minutes). Pull up last week's cabinet agenda. Mark each item C for compression or J for judgment. If yours looks like most, a big share of the meeting goes to C items: updates, report-outs, summaries of things people could have read. Now do the math: minutes spent on C items each week × people in the room × 40 meeting weeks. A cabinet of eight spending 45 minutes a week on report-outs burns 240 leader-hours a year, six full work weeks of senior judgment spent listening to things that could have been a pre-read. Step 2: Flip One Meeting (15 minutes of prep). Move every C item into an AI-organized pre-read with decisions, risks, and open questions labeled. Spend the entire meeting on J items. Watch who talks more when the report-outs disappear. (It's often the person who has been quietly waiting for a real conversation.) Step 3: Seat the Opposition (15 minutes, next cabinet meeting). Pick the decision you're most confident about. Put the opposition prompt on the screen and read the strongest case against your plan out loud as a team. Then ask one question: "Which of these objections can we answer, and which ones can't we?" "We don't have time to use AI more carefully." You're already spending it in the second meeting, where the plan gets rebuilt because nobody could defend the first version. Fifteen minutes with the opposition board is cheaper than one decision you have to take back in public. "My team won't use AI this way. They just want the draft." Of course they do. The draft is visible, and the thinking isn't. So make the thinking visible. Start one meeting a month with a single rule: nobody brings an AI draft until the room has built its own answer first. The Maturity Shift Immature leaders think: "AI makes us faster, so we should use it for everything." Mature leaders think: "AI makes us faster at whatever we already are. We'd better make sure what we are is worth speeding up." Immature leaders ask the tool for the answer. Mature leaders ask it for the landscape, the opposition, and the hardest question in the room, then decide for themselves. Every district and every college can buy the same software. The advantage goes to the cabinet that uses it to sharpen its judgment and widen what it can see. That is how the work stays worthy. Your turn: think about the last major recommendation your cabinet made. How much of the reasoning could the team still explain with the laptop closed? Bonus points if you run the closed-document test in your next meeting and tell us what happened. Most cabinets lack a clear read on the collective judgment underneath all those polished drafts, and whether the team can still think together with the tool turned off. That's the gap THE TEAM INSTITUTE closes. We measure how your cabinet thinks, decides, and perceives as a team, grounded in the TQ research behind IQ × EQ × PQ across 987 leadership teams. Cabinets that closed gaps like this saw performance improve roughly 3x, engagement rise 29%, and outcomes improve 27%, with no increase in burnout. If AI is about to multiply whatever your team already is, would it help to know exactly what that is first? Objectively? Get Started Here with the Team Intelligence Baseline Assessment Help another leader protect their team's thinking before the tools quietly take it over. Repost this with the practice you're trying first. Tag a leader whose cabinet is rolling out AI this fall. Or drop a comment with what the closed-document test turned up. P.S. If you plan a conference or member program for 2027, I built a keynote around that "convince me this is still worthy work" question. It's called The PROOF: higherperformancegroup.com/the-proof Follow DR. JOE HILL and Higher Performance Group