Last month, I sat across a decision-maker...
Who’d just rolled out an AI-powered workflow to the entire department. Adoption was up just three months in, productivity metrics looked super promising, but she was panicking.
“Everyone’s using it,” she told me, “but half my team says they feel less confident than before we started.”
She thought the rollout was failing. I told her it was working.
The Paradox No One Wants to Talk About
ManpowerGroup just released its 2026 Global Talent Barometer, and the headline is startling: AI usage jumped 13%, while worker confidence in technology dropped 18%. For the first time in three years, overall worker confidence declined.
People are using AI more and feeling worse about it.
Nearly nine in ten workers (89%) say they’re confident in the skills they have today. But 43% fear automation will replace their job within two years: up five points from last year. And 64% are clinging to their current employers for stability, a trend called “job hugging.”
The narrative writes itself: AI is breaking the workforce.
But, for me, I think what’s actually happening is an important signal.
The Confidence Dip Is the Learning
If you’ve ever studied a new language, picked up an instrument, or taken on a skill that genuinely challenged you, you know this feeling. At the start, you didn’t know what you didn’t know, and that felt fine. Then you learned enough to realize how much more there was to learn, and your confidence plummeted.
Professionals call this the conscious incompetence stage of learning. It’s the second phase in a well-established model of skill development:
- Unconscious incompetence: You don’t know what you don’t know. Confidence is high.
- Conscious incompetence: You now see the gap. Confidence drops. This is where growth starts.
- Conscious competence: You can do it, with effort. Confidence rebuilds.
- Unconscious competence: It becomes second nature.
What ManpowerGroup’s data is showing us is a global workforce moving from stage one to stage two – at scale. People went from “AI seems manageable” to “I now realize how much I need to learn.” That’s not a collapse. It is actually the beginning of real competence.
Let’s look at the Dunning-Kruger effect: the more you actually know about something, the more accurately you can assess your own limitations. The confidence drop isn’t a sign your team is failing with AI, but instead that they’re finally engaging with it honestly.
The Real Problem Isn't the Confidence Drop - It's What Happens Next
Here’s where I do get concerned. Not about the dip itself, but about what organizations are doing (or not doing) in response to it.
According to ManpowerGroup’s data, 56% of workers received no recent training, and 57% have no access to mentorship.
As Mara Stefan, ManpowerGroup’s VP of Global Insights, put it in Fortune: “Workers are being handed tools without training, context, or support.”
That’s the actual problem. People are hitting conscious incompetence (the moment they most need support) and finding nothing there. No training, mentorship, roadmap for what to do with the discomfort they’re feeling.
It’s like handing someone a piano and walking away. They’ll press a few keys, realize they can’t play, and conclude they never will.
The confidence dip is natural. Leaving people stranded in it is a choice.
What Leaders Can Do Right Now
The good news is that this is a solvable problem. The workforce in a developmental moment that responds directly to the right kind of support. Here’s where to start:
1. Name It Out Loud
Tell your team that the confidence dip is normal. Seriously, just say it. “If you feel less confident about AI now than you did six months ago, that means you’re learning.” When people understand that what they’re experiencing has a name and a trajectory, the anxiety loses its grip.
2. Close the Training Gap With Context, Not Just One-Off Courses
The ManpowerGroup data doesn’t just say people lack training: it says they lack relevant training. Don’t just offer another generic “Introduction to AI” course on your team. Instead, build training around the actual workflows they’re struggling with. What specific tasks are they trying to use AI for? Where are they getting stuck? Start there.
3. Create Peer Learning Spaces
57% of workers have no access to mentorship. But mentorship for AI doesn’t have to mean a formal program. Create spaces where people can share what’s working, what isn’t, and what they’ve figured out. Things like a Slack channel or a 30-minute weekly roundtable. The goal is to normalize the learning process and break the isolation that makes the confidence dip feel personal instead of universal.
4. Separate “Confidence” from “Competence” in Your Metrics
If you’re measuring AI adoption only by usage rates, you’re missing the story. Usage can be high while confidence is low, and that’s exactly what the data shows. Add qualitative check-ins: How comfortable do you feel? Where do you need more support? What would help you feel more confident? These questions surface the real picture.
5. Celebrate the Dip
This might sound counterintuitive, but reward people for identifying what they don’t know yet. When someone says, “I’m not sure I’m using this right,” that’s not a red flag; that’s intellectual honesty. Build a culture where admitting uncertainty is valued. That’s how psychological safety becomes the foundation for sustained adoption.
You're Not Behind. You're Building.
Here’s what I’d tell that leader, and what I’d tell any leader looking at these numbers and feeling uneasy:
A workforce that’s using AI more and feeling less confident about it is a workforce that’s actually engaging with the technology. They’ve moved past the surface. They’re in the messy middle of real learning.
The question isn’t whether the confidence dip will happen – that’s a given. The question is whether you’ll meet your people there, with training, with mentorship, with honesty about the journey, or leave them to figure it out alone.
The organizations that get this right won’t just survive the AI transition. They’ll build teams that are more self-aware, more adaptable, and more resilient than they were before.
The confidence dip where the real story (or transformation) begins.
