The Trust Gap ← Case study
A data story, 2026:

The trust gap

AI use is outrunning confidence in it. What should business leaders do about that?

18,152people surveyed
23markets, Denmark included
8Pioneer markets leading adoption
Each dot ≈ 10 respondents · 23 clusters, one per market · Pioneers in bright blue
Before you start

Would you let AI buy something for you without asking first?

Choose an answer. We'll come back to it at the end.

001 Everyday AI:

How far has AI already reached into daily life?

84%

of people surveyed used AI in the past six months.

So what for leaders

Familiarity, not trust, is driving adoption. Don't read high usage as permission.

002 Assistant to agent:

How far will people let AI go?

Further than you might think. 16% have already used AI that acts without them, in just six months.

Where would you stop?

Select the furthest task you'd hand to AI. Ordered from lower to higher stakes, by our own reading.

Nothing handed over yet.

What people let AI do for them

Share of respondents, %. Switch between the two survey questions.

Source: EY 2026 AI Sentiment Report (n = 18,152, 23 markets)Recreated from the Flourish original
So what for leaders

Customers will delegate small tasks today. Higher-stakes delegation has to be earned.

003 Pioneer markets:

Where is adoption moving fastest?

In eight Pioneer markets, where AI is used more broadly, more often and more deeply in daily life.

The 23 surveyed markets, by stage of AI adoption

Hover or tap a market. Hong Kong SAR and Singapore are shown as dots.

Source: EY 2026 AI Sentiment Report and EY regional releasesRecreated from the Flourish original

Pioneers are ahead on both measures

Share of respondents, %

Pioneer marketsAll 23 markets
Source: EY 2026 AI Sentiment ReportRecreated from the Flourish original
So what for leaders

Pioneer markets show what customers elsewhere will soon expect. Watch them for the early signal.

004 The trust gap:

If people use AI this much, do they trust it?

Not yet. Use is racing ahead, and confidence is lagging behind.

Real or AI? Which of these reviews reads like AI wrote it?

73 in every 100 people

Each dot is 1% of respondents

Source: EY 2026 AI Sentiment ReportNative chart, built in Ceros
73%

fear they won't be able to tell what's real and what's AI-generated.

66%worry about AI systems being hacked or breached
2 in 3say human oversight remains essential
So what for leaders

Every AI interaction is now also a trust test. Plan for the moment it goes wrong.

005 Closing the gap:

What should leaders do about it?

Treat trust as a design problem. The report points to good everyday experiences, reliable data, clear guardrails and accountability.

001Start where trust already exists

Introduce AI agents in frequent, low-stakes moments, like discounts and simple service fixes, where people already welcome help. Let good experiences build the case for more.

002Keep a human in the loop as stakes rise

The higher the stakes, the more people want a person to check, approve or step in. Make that person easy to reach.

003Show your work

Label what AI did, explain why, and let people undo it. Visible, reversible actions are what turn a Maybe into a Yes.

Your answer from the start

You skipped the question at the start. Scroll back up and answer it, and this line will change.