Leaders are sure about AI. Their people are not. That is the AI confidence gap, and new research this week puts a number on it. In a Microsoft survey of UK workers and business leaders, 84% of leaders said they are confident the country can become a world-leading AI economy within five years. Only 41% of employees agreed.

It would be easy to read that as a messaging problem. Leaders believe, employees doubt, so leaders should explain more and sell harder.

I don’t think that is what the gap is telling us.

Confidence is not the same as clarity. And a confidence gap is usually a definition gap wearing a different coat.

Key takeaways

  • New Microsoft research found 84% of UK business leaders are confident about the country’s AI future, compared with 41% of employees.
  • Most employees who use AI at work say it helps them, yet many still worry about job security and about opportunities for young people.
  • A Korn Ferry report says most employers expect entry-level work to change, not disappear, but about half do not yet know what that change looks like.
  • The AI confidence gap closes through clear definitions of work, not more enthusiasm. Leaders have to translate their confidence into specifics.

What’s happening

On October 20 and 21, Microsoft released new UK research at its Frontier Tour event in London. Microsoft’s summary of the research describes two surveys run by Censuswide between October 2 and 12, 2026: one of 5,000 UK employees and one of 1,000 senior business decision-makers.

The headline is the split. Leaders were far more optimistic than employees about where AI is headed. About 90% of leaders said their organization has a clear AI strategy, and 90% expect AI to affect UK society positively over the next five years. Among employees, 49% expected a positive effect on society.

The more interesting finding sits underneath. Among employees who already use AI at work, 71% said it has had a positive impact on their work or career, and 69% said it has improved the quality of their work. Yet 35% of that same group worry AI has lowered their own job security, and 57% are concerned it will reduce opportunities for young people and new entrants to the workforce.

ITPro’s coverage of the Microsoft report, by Nicole Kobie on October 21, framed it as leaders backing the technology while workers are not yet sold. In the release, Microsoft UK & Ireland CEO Darren Hardman pointed to the stakes of that split: “if we remain divided on whether AI will have a positive impact, if we fail to build trust in the technology.”

The same week, on October 20, AMS, a Korn Ferry company, released its Class of 2030 report on early-careers hiring. It surveyed more than 500 CHROs, talent acquisition executives and senior HR leaders, plus more than 5,000 employees and job seekers across 13 countries. Its findings push back on the idea that entry-level work is vanishing: 81% of organizations plan to increase early-career hiring over the next two years, and 76% of talent acquisition leaders say entry-level jobs are not disappearing from their organizations.

But here is the line that stopped me. Nearly three-quarters of organizations expect entry-level work to change rather than be eliminated, and roughly half do not yet know what that change will look like.

Put the two reports next to each other and a pattern shows up. Leaders are confident the future works. Workers are using the tools, seeing real benefits, and still uneasy, especially about what happens to the people just starting out. And the employers who will hire those people say the work will change but have not yet defined how.

Why the AI confidence gap is a leadership problem

In The Next Turn, I describe two changes happening at once. The first turn is the definition of work: what people believe work is for, what it should provide and what it is worth. The second turn is the structure of work: how tasks actually flow, now that AI is becoming part of the infrastructure. This news sits right at the collision.

Look at what employees are telling us. The structure of their work has already changed. Most who use AI say their work is better for it. What has not changed is their understanding of where they stand inside that new structure. If the work gets faster and better, what happens to my role? What happens to the person who would have learned this job the way I did?

Those are not anti-technology questions. They are definition questions. And when leaders answer them with confidence about the destination instead of clarity about the work, the gap widens.

In Chapter 14, The Leader as Translator, I write about credibility over confidence. Leaders earn trust in this moment by naming the tension honestly, defining what matters, explaining tradeoffs and admitting what is still being learned.

This is where credibility becomes more important than confidence.

The Next Turn

A leader saying “AI will be great for us” is confidence. A leader saying “Here is where AI helps in this role, here is what stays fully human, here is how we will judge the work, and here is what we have not figured out yet” is credibility. Employees can test the second one against their own experience. They cannot test the first.

Notice the strategy number, too. About nine in ten leaders say they have a clear AI strategy. A strategy is not the same as a definition. A strategy tells the board where the organization is going. A definition tells the person doing the work what good looks like on Tuesday.

In Chapter 12, Defining Work Again, one of the seven moves is to define AI’s role in the work before employees define it informally. When leaders leave that space empty, people fill it themselves. Some hide their AI use. Some overuse it. Some distrust fast output. And many, as the Microsoft data suggests, quietly worry.

Ambiguity is not neutral.

The Next Turn

The entry-level question is the clearest test

The concern about young people deserves its own attention, because it is where the confidence gap becomes concrete.

The Korn Ferry report describes how junior roles at some leading organizations are already shifting: away from repetitive execution and toward researching and synthesizing information with AI, analyzing data, testing AI-generated outputs, automating workflows and applying human judgment to AI-generated work.

Read that list carefully. Almost every item on it is judgment work. In Chapter 8, The Rise of Judgment, I argue that judgment is moving to the center of value. Testing an AI output means telling plausible from sound. Applying judgment to AI-generated work means knowing the standard and the consequences. Those are skills we used to expect people to build over years of doing the repetitive work first.

So the question is not whether entry-level jobs survive. The question is how people build judgment when the old path for building it has changed. That is a design question, and it belongs to leaders. If roughly half of employers don’t yet know what entry-level work will look like, it is no surprise that workers are uneasy on behalf of the people coming in behind them.

Employees are not wrong to raise it. They are pointing at a gap leaders have not filled yet.

Your next turn

Here are a few moves for this week. None of them require a new strategy. All of them require more specific language.

  • Ask your team the confidence question. Not “Do you like the AI tools?” but “How confident are you about what your role looks like a year from now, and why?” Listen for the gap between benefit and security.
  • Translate one piece of your AI strategy into one role. Pick a single role and write down where AI helps, where it should not be used, and what remains fully human responsibility. Share it and invite corrections.
  • Name what you don’t know yet. If you have not decided how a role will change, say so, and say when you expect to know more. Credibility grows faster from honest uncertainty than from vague optimism.
  • Redesign one entry-level role on purpose. List the tasks AI now handles. Then ask where a new hire will practice judgment, who will review their reasoning, and how they will get feedback on decisions, not just deliverables.
  • Change one question in your one-on-ones. Swap “Did you finish it?” for “How did you decide?” It makes judgment visible and tells people what you value now.

Frequently asked questions

What is the AI confidence gap at work?

It is the difference between how optimistic leaders are about AI and how employees feel about it. Microsoft’s October 2026 UK research found 84% of business leaders confident about the country’s AI future, compared with 41% of employees. Inside organizations, the gap often shows up as leaders talking about strategy while employees wonder what it means for their role.

How can managers build employee trust in AI?

Start with clarity rather than enthusiasm. Define where AI is appropriate in each role, what stays human, and how work will be evaluated now that the process has changed. Be honest about what is still undecided. People tend to trust specific, testable commitments more than broad reassurance.

Is AI eliminating entry-level jobs?

Not according to employers surveyed for Korn Ferry’s Class of 2030 report, released October 20, 2026: 76% of talent acquisition leaders said entry-level jobs are not disappearing, and 81% of organizations plan to increase early-career hiring. Most expect the work itself to change, though about half don’t yet know how. The leadership task is to redesign those roles so new hires can still build judgment.

Closing the gap

Leaders do not need to be less confident about AI. They need to turn that confidence into something people can see and test in their own work.

The structure of work is already changing, and most employees who use AI can feel the benefit. What they are waiting for is a definition of work that tells them where they fit. That is the translator’s job. It is not a memo. It is a series of specific, honest conversations, role by role.

If you want to keep working through these questions, you can find more posts like this on the Next Turn blog.

So here is the question I’d leave you with: if you asked your team today what AI means for their role, would their answer match yours?

Sources