→ Why AI-era confidence is often someone else’s thinking wearing your voice

→ How borrowed certainty fails under pressure

→ One test that separates built judgment from borrowed fluency

There’s a new kind of confidence in the workplace. It sounds informed. It presents well. It survives exactly one follow-up question.

I call it borrowed certainty: conviction assembled from outputs instead of built through reasoning.

It feels identical to earned confidence from the inside. That’s what makes it dangerous.

The research explains the mechanism. A 2025 study on cognitive offloading found AI use correlates with critical thinking at r = –0.68, and offloading itself at –0.75. The more we hand off the thinking, the less thinking capacity we keep.

But the confidence stays. Fluency remains after the reasoning is gone.

The subtle shift

Borrowed certainty shows up in tells:

→ Strong conclusions, vague reasoning

→ Fluency on the what, silence on the why

→ Certainty that collapses the moment the conversation leaves the script

The hidden risk

Leaders running on borrowed certainty make their worst calls at the worst moments: live, under pressure, off-script. Exactly when the borrowed thinking isn’t there to consult.

That’s Cognitive Atrophy’s cruelest feature. It hides until the moment you need what’s missing.

The leadership shift

The question isn’t whether you’re confident. It’s whether you could rebuild the argument from scratch if the tool disappeared.

THE RECONSTRUCTION

Fifteen minutes. Nothing open. No notes, no tabs, no tools. This one is uncomfortable by design, and the discomfort is the measurement.

Set up. Pick one position you hold confidently and have stated publicly in the last month. A strategic call, a recommendation, a view on where your market is going. It has to be something you’d defend in a room.

Close everything. Blank page. Fifteen minutes on a timer.

Rebuild the case from scratch. Write it out as if explaining to someone who disagrees:

1. The claim, in one sentence.

2. The three reasons it’s true.

3. For each reason, the evidence underneath it. Not “the data showed” — what data, from where, how measured.

4. The strongest objection, and your answer to it.

5. What would have to be true for you to be wrong.

Stop at fifteen minutes whether you’re finished or not.

Reading the result

You got through all five. The conviction is built. Whatever role AI played, the reasoning is yours.

You stalled at 3. You have conclusions and you have confidence. What you don’t have is the layer between them. This is the most common outcome and it’s the definition of borrowed certainty.

You stalled at 4 or 5. You’ve never seriously argued the other side, which means you’ve been agreeing with an output rather than testing one.

You couldn’t start. Useful, not embarrassing. It means the position arrived fully formed and you adopted it. Everyone has at least one of these.

The gaps are the data. Not a verdict on you — a map of which convictions are load-bearing and which are decorative. Run it on a different position next month and compare.

Two more this week

• Write your view first, then consult AI. On one meaningful question. Order is everything, and it’s the single highest-leverage habit change in this entire newsletter.

• Invite one hard follow-up. Say to someone whose judgment you trust: push back on my reasoning here. Discomfort is the practice.

Question of the week: Which of my current convictions could I defend without anything the tools gave me?

Before AI reshapes your organization, make sure it isn’t reshaping your ability to think.

The Human Intelligence Assessment is free and takes about ten minutes. It measures the three capabilities AI cannot replace: Awareness, Adaptability, Authenticity.

→ Take the free Human Intelligence Assessment

Lead what AI can’t.
JoshHMiller.com

→ Why the AI era is producing more output and less understanding

→ What leaders stop hearing when everything gets summarized

→ One listening discipline no tool can replicate

We have never produced more communication. And we may have never listened less.

Every message summarized. Every meeting recapped. Every conversation compressed to bullet points before a human ever sits with it.

Summaries capture what was said. They lose what was meant.

The research on coaching cultures shows what listening is worth: in ICF and Human Capital Institute studies, 51% of organizations with strong coaching cultures report higher revenue than industry peers, and 62% of their employees describe themselves as highly engaged.

And the foundation of every coaching culture is the same unglamorous skill. Leaders who actually listen.

The subtle shift

When AI mediates communication, leaders increasingly consume their people secondhand:

→ The recap instead of the meeting

→ The summary instead of the struggle

→ The sentiment score instead of the sentiment

The hidden risk

Hesitation doesn’t survive summarization. Neither does doubt, energy, or the thing someone almost said.

Leaders who listen only to processed input slowly lose contact with the organization they’re supposedly leading.

The leadership shift

In a workplace where everyone has infinite generation capacity, the scarce resource is genuine reception. The leader who fully receives what people are saying, and not saying, holds an advantage no tool can copy.

THE SECONDHAND AUDIT

Five statements about how information reaches you. Score each 1 to 5, where 1 is rarely and 5 is most of the time. Three minutes.

1. I read the recap more often than I attend the meeting.

2. I have formed a view about someone’s performance largely from written summaries of their work.

3. My sense of how the team is doing comes more from surveys, dashboards, or sentiment tools than from conversations.

4. I can’t remember the last time someone told me something in person that genuinely surprised me.

5. Concerns reach me already framed, already sequenced, and often already resolved.

Your score

• 5 to 10 — Firsthand. You’re still in direct contact with your organization. This erodes by convenience, so notice when a recap starts replacing a room.

• 11 to 17 — Mediated. You’re getting an accurate picture and an incomplete one. What’s missing isn’t facts. It’s hesitation, energy, and the thing someone almost said.

• 18 to 25 — Fully secondhand. You’re leading a version of your organization that has been summarized for you. Everything reaching you is true. It’s also processed, and you have no way to inspect what was filtered out, because filtering is what processing is.

What summaries remove

→ Hesitation. Nobody transcribes a pause. Pauses are where the real answer usually lives.

→ Energy. A recap can’t tell you whether the team was resigned or convinced.

→ The unfinished thought. The thing someone started to say and stopped. That never makes it into a summary because it never became a sentence.

Three moves this week

1. Take one meeting with no recap, no transcript, no notes app. Just attention. Afterward, write down one thing you caught that wouldn’t have survived a summary. If you catch nothing, run it again with a harder meeting.

2. Reflect before responding. In your next 1:1, say: what I’m hearing underneath this is... Then stop and let them correct you. The correction is the information.

3. Listen for the unsaid. Ask once this week: what are you not telling me because it isn’t fully formed yet? Then wait longer than is comfortable. The urge to fill that silence is the same urge that reaches for the tool.

Question of the week: When did I last hear someone’s actual words instead of a summary of them?

If this landed, the most useful thing you can do is send it to the one person on your team who would recognize themselves in it.

→ Restack it so other readers here find it

→ Forward it to the person it is actually about

Lead what AI can’t.
JoshHMiller.com

→ Why employees feel processed instead of seen

→ The disconnect between what leaders are rewarded for and what people need

→ One habit that makes recognition specific enough to matter

Here’s a finding that should stop every executive mid-scroll.

Hogan’s research across roughly 10,000 employees and 21,000+ executives in 25 countries found zero overlap between the traits executives are rewarded for and the traits employees actually want in their leaders.

Zero. Not small. Zero.

Organizations promote for drive, ambition, and results. Employees want integrity, communication, and someone who sees them.

And AI is widening the gap, because AI is spectacular at the rewarded traits and useless at the wanted ones.

The subtle shift

As AI handles more of the measurable, people increasingly evaluate leaders on the immeasurable:

→ Did you notice my contribution, or just my output?

→ Do you know what I’m trying to become, or just what I produce?

The hidden risk

Generic recognition is now instantly detectable. Everyone knows what AI-drafted praise sounds like: polished, warm-ish, and completely interchangeable.

Recognition that could apply to anyone lands as recognition of no one.

The leadership shift

Being seen requires evidence of attention. Specificity is the proof. And specificity requires the one thing you can’t delegate: actually paying attention.

THE SPECIFICITY TEST

Find the last piece of recognition you gave. A message, an email, a line in a review, something you said in a meeting. Pull up the actual words. Three tests. Three minutes.

TEST 1 — The swap test

Replace the person’s name with anyone else on your team. Does the sentence still make sense?

If it does, you recognized a role, not a person. That still counts for something. It just doesn’t do the thing you think it does, because the person receiving it can run this same test in half a second and usually does.

TEST 2 — The room test

Could someone who wasn’t there have written this?

If yes, you recognized an outcome you heard about, not work you witnessed. Outcomes are visible from a distance, which is precisely why recognizing them proves nothing about your attention.

This is the test AI-drafted praise fails hardest. A tool can only ever write from a distance.

TEST 3 — The how test

Does your recognition name what they did, or what happened because of what they did?

“Great results on the Q3 push” is an outcome. “The way you reframed that objection in the client meeting changed the whole room” is a method.

Method is the proof of attention. It’s also the only kind of recognition that tells someone what to repeat.

The rewrite

Take the recognition that failed one or more tests and rewrite it in one sentence, naming one specific choice that person made. Not what happened. What they decided to do.

If you can’t name a specific choice, you’ve found something more useful than a rewrite. You’ve found a person whose work you’ve been evaluating from the outputs.

Two more this week

1. Write it yourself. Unpolished and specific beats polished and generic, and the gap between those two is now instantly detectable by anyone who has used the same tools you have.

2. Ask one person: what work are you proudest of that nobody noticed? Then notice it. The question is only worth asking if you do the second part.

Question of the week: Could the recognition I gave last week have been written by someone who wasn’t in the room?

Before AI reshapes your organization, make sure it isn’t reshaping your ability to think.

The Human Intelligence Assessment is free and takes about ten minutes. It measures the three capabilities AI cannot replace: Awareness, Adaptability, Authenticity.

→ Take the free Human Intelligence Assessment

Lead what AI can’t.
JoshHMiller.com

→ Why the value of speed just collapsed

→ What deliberate slowness signals to a team

→ One pause that upgrades every decision it touches

For your entire career, speed was an advantage. Fast answers. Fast decisions. Fast replies.

Then AI made speed free.

And anything free stops being an advantage.

Here’s what the research shows about the cost of frictionless speed. In an MIT study covered by TIME, ChatGPT users showed the lowest brain engagement of any group, with effort declining over time. 83% of heavy users couldn’t recall what they had just written minutes later.

Speed without engagement isn’t thinking. It’s throughput.

The subtle shift

When everyone moves at machine speed, pace stops differentiating anyone. What differentiates now is what speed can’t produce:

→ The question nobody thought to ask

→ The risk everyone else skimmed past

→ The judgment call that required sitting with the problem

The hidden risk

Leaders who match the machine’s tempo teach their teams that reflection is a luxury.

Reflection is not a luxury. It’s where judgment is manufactured. Remove it and you get Cognitive Atrophy on an organizational schedule.

The leadership shift

Slowness used to signal indecision. In the AI era, deliberate slowness signals something else entirely: this decision matters enough to think about.

THE STAKES CLOCK

AI can give you the answer in four seconds. The question is how long the commitment should take. Two questions place any decision. Two minutes.

1. Can it be undone?

2. How far does it travel if it’s wrong?

Those two answers place any decision in one of four tiers.

TIER 1 — Reversible, contained A draft, a scheduling call, an internal format change.

• Dwell time: none. Decide at tool speed. This is what AI is for, and hesitating here turns deliberateness into a tax on your team.

TIER 2 — Reversible, travels A process change, a team-wide communication, a shift in how you run something.

• Dwell time: one night. Not because you need more information. Because you need to see it again without the momentum that produced it.

TIER 3 — Hard to undo, contained A hire, a role change, a difficult conversation, a commitment to one client.

• Dwell time: 24 hours, plus one dissenting voice. Name the person most likely to disagree and ask them before you commit. If you can’t think of anyone who would push back, that’s the finding, not the absence of one.

TIER 4 — Hard to undo, travels Strategy, restructuring, a public position, anything that changes what people believe about the organization.

• Dwell time: a week minimum. And before you commit, write down the single piece of evidence that would change your mind. If nothing would, you are not deciding. You are announcing.

The test

Look at your last five significant decisions. Name the tier for each, then name how long you actually spent.

The gap is where the tool set your pace instead of the stakes.

Most leaders find it in Tier 3. Tier 4 decisions still get respected because they feel big. Tier 3 decisions feel manageable, which is exactly why they get made at the speed of whatever produced the recommendation.

Three ways to build the dwell time in

1. Institute a 24-hour rule for one class of decision. The answer can be instant. The commitment shouldn’t be.

2. Open one meeting with two minutes of silent thinking before anyone speaks. Watch what changes about what gets said.

3. Block thirty minutes weekly for unassisted reflection. No inputs. Just you and the hardest problem on your plate.

Question of the week: Which recent decision did I make at the speed of the tool instead of the speed of the stakes?

If this landed, the most useful thing you can do is send it to the one person on your team who would recognize themselves in it.

Forwarding beats every growth tactic I have tried, and it is the only one that arrives with trust already attached.

→ Restack it so other readers here find it

→ Forward it to the person it is actually about

Lead what AI can’t.
JoshHMiller.com

→ Why AI adoption is rising while leadership confidence isn’t

→ The trust transfer leaders are attempting, and why it fails

→ One habit that keeps trust anchored where it belongs

Something odd is happening in organizations rolling out AI. The tools get more capable every quarter. And trust in leadership doesn’t move at all.

Russell Reynolds found that 82% of leaders say AI understanding will be mandatory for the C-suite, yet only 41% feel personally confident in their own. And 54% worry generative AI will stunt critical thinking.

Read those together: leaders are being asked to project confidence about something most of them privately don’t feel.

So some are borrowing it. Presenting AI’s analysis as certainty. Hiding behind the output.

The subtle shift

Teams can tell the difference between a leader who used AI and a leader who’s hiding behind it.

→ Used it: “Here’s the analysis, here’s where I disagree with it, here’s my call.”

→ Hiding: “The data says.” “The model recommends.” “The analysis is clear.”

One owns the decision. The other launders it.

The hidden risk

Every time a leader attributes a decision to the tool, they teach the team that accountability lives somewhere else.

Trust doesn’t transfer to the AI. It just evaporates.

The leadership shift

Your people don’t need you to be more certain than the machine. They need you to be more accountable than it.

That’s a bar AI can never clear.

THE ATTRIBUTION AUDIT

This one doesn’t run on how you think you communicate. Pull up the last three significant decisions you announced — the actual Slack message, email, or meeting notes. Read your own words. Then score. Four minutes.

Score each 1 to 5, where 1 is never and 5 is most of the time.

1. My decisions arrive attached to what the analysis showed rather than to what I concluded.

2. I have described an output as clear, obvious, or conclusive without saying whether I agreed with it.

3. When a decision went badly, my explanation included where the information came from.

4. I cannot recall the last time I told my team I disagreed with something a tool produced.

5. I would find it harder to defend a decision if the analysis behind it were removed from the room.

Your score

• 5 to 10 — Anchored. Your judgment is visible in how you communicate. Keep saying the disagreement part out loud, because that’s what makes it visible.

• 11 to 17 — Drifting. You are owning your decisions and describing them as though the analysis owned them. The gap is in the language, not the thinking, which means it’s fixable this week.

• 18 to 25 — Laundering. Accountability has moved into the passive voice. Your team is hearing a system with a person attached rather than a person with a system attached. They will not say this to you.

The swap

Every laundered sentence has an owned version. The difference is one clause.

Instead ofSayThe data says we should consolidate.I decided to consolidate. Here’s what convinced me.The model recommends this candidate.I’m going with this candidate. The analysis flagged something I’d missed.The analysis is clear on timing.I think the timing is right, and here’s where I’d push back on my own reasoning.

The owned version is longer. That’s not a defect. The extra clause is the accountability, and removing it is exactly how the laundering happens.

Three moves this week

1. Replace “the data says” with “I decided, and here’s why” in your next three decisions. Count how often you catch yourself reaching for the first one.

2. Name your disagreement with an output publicly, at least once. Not the whole reasoning. Just: the analysis said X, I’m doing Y. Visible judgment is the only kind that builds trust.

3. When you don’t know, say so. Borrowed certainty reads as exactly what it is, and your team has been reading it longer than you think.

Question of the week: In my last big decision, did my team hear my judgment or the tool’s?

Before AI reshapes your organization, make sure it isn’t reshaping your ability to think.

The Human Intelligence Assessment is free and takes about ten minutes. It measures the three capabilities AI cannot replace: Awareness, Adaptability, Authenticity.

→ Take the free Human Intelligence Assessment

Lead what AI can’t.
JoshHMiller.com

→ Why AI reduces the need to give answers

→ How coaching becomes a critical leadership skill

→ One check to run before your next conversation

AI is exceptional at giving answers.

Which means leaders don’t need to.

But that doesn’t reduce the value of leadership. It redefines it.

The real shift

When answers are abundant, thinking becomes the scarce resource.

According to the International Coaching Federation, organizations with strong coaching cultures see higher leadership effectiveness, engagement, and performance.

That’s not surprising. Because coaching does something AI can’t. It develops how people think, not just what they know.

The leadership risk

If leaders continue to default to providing answers, especially in an AI-enabled environment, they unintentionally reinforce dependency, surface-level thinking, and low ownership.

Over time, this contributes to Cognitive Atrophy. People rely more on outputs and less on their own reasoning.

The leadership evolution

The best leaders won’t compete with AI on answers.

They’ll differentiate on curiosity, inquiry, and cognitive development.

That’s the Human Intelligence Advantage in action.

THE DEVELOPMENT CHECK

Run this before your next conversation with someone who is stuck. Three questions, ninety seconds.

1. Whose problem am I about to solve?

• If you already know what you’re going to tell them, you are about to close a loop rather than open one. That’s sometimes correct. Decide it deliberately rather than by reflex.

2. What will they be able to do next time because of this conversation?

• If the honest answer is “come back to me,” you are building dependency. If it’s “run that reasoning themselves,” you are building capability.

3. What am I optimizing for right now — speed or growth?

• Both are legitimate. A live incident wants speed. A recurring pattern wants growth. Most leaders default to speed and call it efficiency, then wonder why the same problem returns.

The two questions that replace the answer

Open with: What’s your initial thinking? Not “what do you think,” which invites a conclusion. “Initial thinking” gives permission to be unfinished.

Follow with: What assumptions are you making? This is the one that does the work. It moves them from their answer to their reasoning, which is what transfers.

The discipline underneath both: Ask, then stop talking. The urge to fill the silence is the same urge that reaches for the tool.

Question of the week: Where might I be solving instead of developing?

Lead what AI can’t.

→ Why constant availability is not leadership presence

→ How AI accelerates responsiveness expectations

→ One rep that restores meaningful presence

Leaders today are more available than ever.

Slack. Email. AI-assisted replies. Always on. Always reachable.

But here’s the uncomfortable truth: responsiveness is increasing while presence is declining.

According to Gallup, employees who feel genuinely seen and supported by their managers are significantly more engaged, regardless of how often they communicate.

What’s actually happening

AI is optimizing for speed, clarity, and efficiency.

But leadership presence is built on attention, emotional attunement, and depth of engagement.

Those don’t scale through faster replies.

The hidden tradeoff

Leaders are substituting quick responses for meaningful interaction, and availability for connection.

Over time, this creates a subtle form of Cognitive Atrophy. Attention gets fragmented. Depth gets reduced. Connection gets diluted.

The leadership shift

Presence is no longer about being accessible. It’s about being fully engaged when it matters most.

That’s a choice, not a system.

THE PRESENCE REP

One week. One conversation. No tools.

Choose the conversation. Look at your calendar and find the one conversation this week where the person on the other side would most benefit from your full attention. Usually it’s the one you’ve been handling in messages because it’s faster.

Move it out of text. Call, walk, or sit down. Fifteen minutes is enough. The channel change is most of the rep.

Run it under three constraints:

1. No device in your hand or on the table. Not face down. Away.

2. No drafted response. Whatever you say is composed in the moment, in front of them.

3. One question you don’t know the answer to. Ask it and let the silence sit. Do not fill it.

Afterward, answer one thing: what did you learn that would not have surfaced in writing?

If the answer is nothing, run it again with someone else. If the answer is something, that gap is what fast replies have been costing you.

The tell to watch for. If you find yourself mentally drafting a reply while they’re still talking, that’s the atrophy. Attention is a capability, and it weakens without use like any other.

Question of the week: Where might I be responsive but not truly present?

Lead what AI can’t.

→ How AI changes the risk of being wrong

→ Why people may speak less, not more

→ One audit that surfaces what your team is holding back

AI creates a new kind of pressure.

When answers feel instantly available, and often correct, being wrong starts to feel more visible. More avoidable. More costly.

According to Google’s Project Aristotle research, psychological safety is the number one predictor of high-performing teams.

But here’s what’s changing: the definition of risk is shifting.

The subtle behavioral shift

When the right answer feels accessible, people hesitate to share incomplete ideas, early thinking, and unstructured perspectives.

Not because they lack insight. Because AI has raised the perceived bar for contribution.

And when that happens, thinking moves from out loud to internal. Collaboration becomes filtered instead of exploratory.

The hidden risk

Leaders lose access to raw thinking.

They only hear polished ideas, refined conclusions, post-processed input.

That’s Cognitive Atrophy at the team level. Less visible thinking, less shared intelligence.

The leadership shift

Psychological safety is no longer just about permission to speak.

It’s about permission to think in public again.

That’s a different leadership muscle.

THE UNFINISHED THINKING AUDIT

Five statements about your team, not about you. Score each 1 to 5, where 1 is rarely and 5 is regularly. Three minutes.

1. Ideas arrive in my meetings already structured, formatted, and complete.

2. I can’t remember the last time someone said “I’m still working this out” in front of the group.

3. Written updates from my team read more polished than the people who wrote them sound in conversation.

4. Disagreement, when it comes, comes privately and afterward rather than in the room.

5. I rarely hear a half-formed idea that turns into a better one during the discussion.

Your score

5 to 10 — Thinking is still visible. Protect it. This erodes quietly.

11 to 17 — Filtering has started. You are getting conclusions rather than reasoning. The intelligence is still there; it just isn’t reaching you.

18 to 25 — Fully post-processed. You are seeing output, not thinking. Every idea reaching you has already survived a private filter you cannot inspect.

The three moves

• Reward unfinished thinking. Say: talk me through your early thoughts.

• Model intellectual vulnerability. Share what you’re still figuring out, not just what you know.

• Normalize iteration over accuracy. Insight comes from progression, not perfection.

And one thing to say out loud this week: I want the version before you cleaned it up. Say it in a meeting, not in a message. It only works if people watch you mean it.

Question of the week: Where might my team be holding back ideas because they’re not fully formed?

Lead what AI can’t.

→ Why AI is reshaping where people look for answers

→ How authority shifts without leaders noticing

→ One behavior that restores leadership credibility

AI is becoming the fastest voice in the room.

Instant answers. Confident responses. Always available.

And here’s the shift most leaders haven’t fully processed: speed is starting to outrank trust.

According to Microsoft, a growing percentage of knowledge workers now rely on AI for guidance, often before seeking input from colleagues or managers.

That’s not just a behavior change. It’s an authority transfer.

What’s really happening

Leaders aren’t losing credibility because they know less.

They’re losing it because they’re no longer the first place people go to make sense of things.

And when that happens consistently, something subtle begins to erode. Context disappears. Judgment gets outsourced. Interpretation gets flattened.

This is where Cognitive Atrophy shows up at the organizational level. Not just individuals thinking less, but teams defaulting to external intelligence over internal leadership.

The new leadership standard

Authority is no longer built on having answers. It’s built on making meaning from answers.

Anyone can generate information. Very few can contextualize it, challenge it, and translate it into action.

That’s the Human Intelligence Advantage.

Strengthening your Human Intelligence Advantage

• Become the interpreter, not the source. Ask: What does this actually mean for us?

• Add context AI can’t see. Culture, history, risk, nuance. This is where authority is built.

• Show your thinking, not just your answer. Trust grows when people see how you think, not just what you conclude.

THE AUTHORITY LOCATOR

Four phases. One question each. Answer honestly about your own team, then find where you are. Two minutes.

• PHASE 1 — SOURCE When someone on your team hits an ambiguous problem, do they bring it to you first?

  • You are still the first stop. Authority is intact. The risk here is complacency, not erosion.

• PHASE 2 — SECOND OPINION Do they arrive with an answer already formed and ask you to confirm it?

  • The order has flipped. You are validating rather than interpreting. Most leaders are here and don’t know it, because the meetings still happen and still feel productive.

• PHASE 3 — ESCALATION ONLY Do they come to you mainly when something has already gone wrong?

  • You have been routed out of the thinking and into the cleanup. Your visible work is now correction, which is the least authority-building work there is.

• PHASE 4 — BYPASS Have you learned about a significant decision after it was made, more than once this quarter?

  • Authority has transferred. Not maliciously. Your team found something faster and nothing in the system told them not to.

Where you are is less important than which direction you are moving. Ask a trusted report which phase they would put you in. Their answer is the real one.

Question of the week: Where might my team be turning elsewhere before turning to me?

Lead what AI can’t.

→ Why AI makes it easier to avoid ownership

→ How responsibility subtly shifts to systems

→ One check that restores accountability

AI doesn’t just support decisions.

It quietly absorbs ownership.

Recommendations. Rankings. Predictions.

And with each one, leaders are given a subtle narrative: the system suggested it.

According to the World Economic Forum, organizations are already struggling to define accountability in AI-influenced decisions.

That’s the signal.

Responsibility isn’t disappearing. It’s becoming diffused.

The real leadership risk

When decisions are influenced by AI, leaders begin to psychologically distance themselves from outcomes. Not intentionally. But progressively.

This is another form of Cognitive Atrophy. Not thinking less, but owning less.

And leadership without ownership is just coordination.

Strengthening your Human Intelligence Advantage

Name the decision owner, explicitly. Even when AI is used, someone owns the outcome.

Replace “the system recommended” with “I decided.” Language shapes accountability.

Define non-delegable decisions. What will you never outsource, no matter how good the system gets?

THE OWNERSHIP CHECK

Pick one real decision you made in the last two weeks where AI was involved. Not a hypothetical. Run these three questions on it. Three minutes.

1. The defense test. If this decision goes badly and you have to defend it in a room, what do you say?

   • If your answer contains the word recommended, suggested, indicated, or showed, you have not fully taken the decision. You have countersigned one.

2. The subtraction test. Remove the AI output entirely. Same information, no recommendation. Would you have reached the same conclusion?

   • If yes, the tool sharpened you. If no, the tool decided and you approved. Both are legitimate. Only one is leadership, and you should know which one you did.

3. The naming test. Say out loud, to yourself: I decided this. Then name the one assumption it rests on.

   • If the assumption came from the output rather than from you, that is the assumption to test before the next decision like it.

Then define one non-delegable. Write down one category of decision you will never hand to a system, however good it gets. Say it to your team this week. A boundary nobody knows about isn’t a boundary.

Question of the week: Where might I be using AI to justify decisions instead of owning them?

Lead what AI can’t.

That’s the part we’re not talking about.

Because the real risk of AI isn’t misinformation. It’s dopamine.

Neuroscientist Andrew Huberman explains that intermittent rewards — unpredictable wins — create powerful dopamine loops that reinforce behavior. It’s the same mechanism used in gambling, social media, and now AI.

Sometimes AI gives a brilliant answer. Sometimes it’s wrong but sounds confident. Sometimes it surprises you.

That unpredictability is the hook.

And leaders feel it.

HOW THE TRAP SHOWS UP IN LEADERSHIP

1. Dopamine leadership. Speed and certainty start to feel productive, even when judgment erodes.

2. Confidence contagion. When leaders stop questioning AI, teams stop questioning anything.

3. Machine-tempo culture. Fast Company warns that AI-accelerated work expectations are already contributing to burnout and decision fatigue. Humans are expected to operate at machine speed.

This is where Human Intelligence matters most.

Buzzword: AI as a thinking partner

Reframe: It can quietly become a thinking replacement

Action: Treat AI output as a stimulus, not a decision

AI can scale execution. Only humans can scale judgment.

THE PULL AUDIT

Five statements. Score each 1 to 5, where 1 is never and 5 is most days. Two minutes.

1. I have opened an AI tool before deciding what I actually wanted from it.

2. I have accepted an output faster because it arrived quickly, not because it was better.

3. I have gone back to re-prompt for a more satisfying answer rather than working with what I had.

4. I would struggle to reconstruct the reasoning behind something I produced with AI this week.

5. I notice a small lift when the output is unexpectedly good.

Your score

• 5 to 10 — Deliberate use. You are using the tool. Keep the pause in place.

• 11 to 17 — Early pull. The loop is forming. It has not yet reached your decisions, but it has reached your habits.

• 18 to 25 — Conditioned. Speed and certainty have started to feel like quality. The three questions below are not optional this week.

The practice, whatever you scored. Before acting on any AI output today:

→ What am I assuming is true because it sounds certain?

→ What evidence would change my mind?

→ What’s the human cost if this is wrong?

Question of the week: Where is the pull strongest for me — speed, certainty, or relief?

Lead what AI can’t.