By Sloane, Marketing, HUMAN HIVE
Do managers trust AI more than their own judgment? About a third of them say yes: in Charter's AI Pulse survey of 5,002 U.S. AI users, 34.8% of managers and 32.8% of executives trust AI over their own judgment on work decisions, against 16.3% of individual workers. The fix is a human one: name the decisions that always get a human check, teach people how to verify, and keep the experienced voices in the room.
The 60-second version
Managers and executives are about twice as likely as workers to trust AI over their own judgment.
The top reason people use AI output they doubt is time pressure, not trust.
This week, name the three decisions on your team that always get a human check before anyone acts.
What happened
On November 16, 2026, Fortune put a sharp question to leaders: everyone worries that Gen Z trusts AI too much, so why are executives the ones handing it their judgment? The numbers come from Charter's AI Pulse report (Fall 2026), a survey of 5,002 full-time U.S. desk workers who use AI on the job, fielded in August 2026 and weighted to the U.S. Census workforce population. Axios first reported the findings on October 21.
From the full report:
24.9% of all AI users say they trust AI more than their own judgment on work decisions. That rises to 34.8% of managers and 32.8% of executives, against 16.3% of individual workers.
About a third of managers and executives trust AI more than a knowledgeable colleague. Fewer than 16% of workers say the same.
81% of executives say they always or usually fact-check AI output, compared with 71% of workers and 68% of managers.
Yet executives are 3.3 times more likely than workers to use AI output they have doubts about. Across everyone, 63.6% have used an output they doubted.
Brian Elliott, Charter's executive-in-residence, told Fortune: "We surveyed over 5,000 AI users at work and found sharper divides by career stage than by age."
Why it happened: the human side
This is not a story about gullible bosses. Charter asked people why they used AI work without verifying it. The top reasons were being short on time (43.7%), not knowing how to verify the work (38.7%) and deciding the stakes were low (34.9%). Executives were the most likely to say the stakes were low (42%).
Read that list again. Two of the three are human habits, and the third is a skill nobody taught. Leaders are busy, so they take the fast answer. Nobody showed them a quick way to check it. And "low stakes" is a judgment call made by the very person who is short on time.
There is a second shift underneath. When Charter asked where people turn first for ideas or help thinking something through, coworkers still came first overall (29.5%). Among daily AI users the order flips: 27.5% turn to AI first and 23.1% to a coworker. When the machine becomes the first stop, the experienced person down the hall stops getting asked. That is how context goes missing.
What Mitch has taught on this
At the Central Valley SHRM conference in October, Mitch told a story from a training day at a natural gas company. A team was standing near a wellhead, running a pretend repair. A worker with about five years in the field read the fix off his phone: "ChatGPT says we should do this." Off to the side stood Chuck, 30 years in, arms crossed.
In Mitch's telling, Chuck said: "Well, what you didn't tell that machine was that it was 42 degrees this morning, and it froze overnight. You didn't tell him it's gonna be 80 by the afternoon." Then: "If you do what he said, we're all gonna die out here."
Mitch's line to the room: "That's wisdom." And then: "So wisdom in the future is the key. It's the key to all the human stuff because we know what things feel like. We know what the stakes are in the game. We know what failure feels like. AI does not know any of this."
The machine wasn't wrong about what it was given. Nobody gave it the frost. That is the job of human judgment, and it gets harder to do the more you hand the thinking over. In the same talk, Mitch put the leadership task plainly: "We're gonna have to learn where our discernment comes in."
Watch how this plays out live in the Decision Heist official trailer. Decision Heist shows people that their decisions still matter and their discernment still matters.
Best practices for leaders, L&D and sales managers
Write the high-stakes list. The team lead writes down the decisions that always need a human check before anyone acts: money over a set amount, anything about a person's job, safety, legal, and customer promises. Post it where the team works. Charter recommends this too: define "high stakes" so time pressure stops making the call.
Budget the check, not just the task. If the top reason for skipping verification is time, put the time in the plan. A 10-minute review slot on any AI-assisted decision on the list costs less than the cleanup.
Teach three ways to verify. L&D can run a 30-minute session: confirm the source, check the underlying numbers, and ask a teammate who knows the territory. 38.7% of people who used a doubtful output said they didn't know how to check it. That gap is trainable.
Ask "what didn't we tell it?" Before acting on an AI recommendation, the decision owner asks the room one question: what does this tool not know about our situation? That is Chuck's question, and anyone can learn to ask it.
Leaders go first, out loud. Sales managers and executives say in the meeting how they checked an AI answer before using it. When the boss models verification, the team copies it.
What's driving people crazy this week
Workers keep describing bosses who route real decisions through a chatbot. Futurism reported on a legal tech employee whose boss made "structural company decisions based solely on his conversations with ChatGPT," including personnel choices. Her duties changed three times in a few months, each time after a new chatbot conversation, and she quit. The human fix is simple and hard: AI can inform the decision, but a named person owns it, explains it, and can say what the tool didn't know.
What this means for the next 1 to 3 years
The "context brief" becomes a core leadership skill. The leaders who get the most from AI will be the ones who are best at telling it what it doesn't know: the frost, the client history, the politics. Expect that skill to show up in leadership programs and job descriptions.
Veterans become the verification layer. Smart teams will pair their most experienced people with their heaviest AI users, on purpose, so the Chucks are in the room when the phone comes out. Charter found nearly one in five AI users (19.9%) say AI is making it harder for early-career workers to build skills, so that pairing helps both sides.
Decision ownership gets written down. As agents start acting, not just suggesting, companies will name a human owner for each kind of consequential decision, with the authority to stop the process. The firms that do it early will move faster, because their people will know when they're allowed to trust the tool.
How we evolve
The human skill to build next is discernment: knowing what a situation needs that the data doesn't show. It matters more as AI gets better, because a confident answer is easier to accept than a hard question. Mitch's New Rules of Work start with this one: "Knowledge alone is no longer the advantage." What you do with knowledge is.
One small daily practice: before you ask AI for a recommendation on a real decision, write down your own answer in one line. Then compare. Where you differ, ask why. You keep your judgment in shape, and you learn what the tool is good at.
Try this today
Ask at your next team meeting (5 minutes): "When did AI give us an answer this month that we used without checking?" Listen without blame.
Draft your high-stakes list (15 minutes): three to five decision types that always get a human check, with an owner for each.
Run Chuck's question once (5 minutes): on the next AI-assisted decision, ask "What didn't we tell it?" before anyone acts.
Free tool: the 3-line decision check
Copy this into your team notes: "Before we act on an AI recommendation for a decision on our high-stakes list, the owner answers three lines: 1. What did we not tell the tool about our situation? 2. How did we check the key fact or number? 3. Who with experience in this area looked at it? If any line is blank, we pause."
FAQ
Do executives trust AI more than their own judgment?
About a third say so. In Charter's AI Pulse survey of 5,002 U.S. AI users (Fall 2026), 32.8% of executives and 34.8% of managers said they trust AI more than their own judgment on work decisions, compared with 16.3% of individual workers.
Why do people use AI output they doubt?
The top reasons in the same survey were being short on time (43.7%), not knowing how to verify the work (38.7%) and deciding the stakes were low (34.9%).
How can a team keep human judgment in AI decisions?
Name the decisions that always need a human check, give people time and a simple method to verify, and make sure an experienced person reviews high-stakes calls. Ask "what didn't we tell it?" before acting.
Where HIVE can help
Leaders decide with half the facts every day. Practice it together, live, before it counts. Rolling out AI and not sure your people know when to check it? See our AI enablement work. Related: why senior leaders win with AI and junior staff fall behind.