The promise of AI in the office was supposed to arrive fast. It didn’t. A new survey finds that despite years of hype and public interest, the actual use of AI in daily work remains slow, informal and uneven.
The Konstanz AI study, now in its third wave, surveyed 1,105 employees in May 2026. It found that the share of workers using AI in their jobs rose only slightly, from 35% to 38%, compared with the previous year. That is a small gain over a period when the technology has dominated headlines and conferences.
Phys.org’s survey suggests that the gap between what people talk about and what people actually do at their desks is wider than anyone admitted. The results raise questions about whether the AI revolution is truly happening at the pace the press has suggested.
Survey Background and the Key Numbers
The Konstanz AI study is a recurring survey of employee attitudes toward artificial intelligence in the workplace. The third wave, conducted in May 2026, asked 1,105 employees about their use of AI tools.
The key comparison is simple: last year, 35% of respondents said they used AI in their work. This year, that figure rose to 38%. The difference is three percentage points.
That is not nothing. Three points is a real change in a sample of this size. But it is also a modest one, especially against the backdrop of a year in which AI seemed to be everywhere. The question the survey raises is why the adoption number moved so little while the conversation around AI grew so loud.
What the Survey Measures
The survey asks employees whether they use AI in their day-to-day work. The question is straightforward: do you use AI tools as part of your job?
The distinction between “public interest” and “day-to-day work” is crucial here. People can be fascinated by a technology without using it. They can read about it, watch videos about it, and talk about it at dinner parties. None of that shows up in a measure of actual adoption.
What the survey captures is the gap between what people consume and what they apply. It is a measure of practice, though the source does not explicitly characterize it as such.
Adoption Remains Slow and Informal
One of the most striking findings is that AI adoption remains slow and informal. The survey’s description of the situation — slow, informal and uneven — captures the texture of the current moment.
Informal use means that employees are experimenting with AI tools on their own, outside of formal training programs or company directives. They are downloading apps, signing up for trials, and figuring out how to apply machine learning to their tasks. This is the kind of adoption that happens when people find value on their own, rather than waiting for permission.
But the survey does not describe the instability of informal use. It simply reports that adoption is slow and uneven.
Practical Implications for Managers
The survey has practical implications for managers and HR professionals. If AI adoption is slow and informal, then the tools that companies invest in may not be used as intended.
Here are a few steps companies can take to bridge the gap between interest and practice:
- Training Programs: Offer regular training sessions to help employees develop the skills they need to use AI tools effectively.
- Clear Communication: Explain how AI can improve specific workflows, so employees understand the benefits.
- Pilot Projects: Run small-scale pilot projects to demonstrate the value of AI in real-world settings.
- Support Systems: Provide ongoing support, such as helpdesks or dedicated teams, to address questions and issues that arise during implementation.
- Incentives: Consider offering incentives for employees who successfully adopt AI tools.
None of these steps guarantee adoption. But they move the locus of control from individual curiosity to institutional commitment.
The Public Interest Question
The survey also raises a broader question about how society measures technological progress. When we talk about AI adoption, we often focus on metrics like the number of users or the amount of data processed. Those are important numbers, but they do not capture the full picture.
The survey shows that public interest in AI is high. But that interest is not translating into routine use. The gap between the two is the story here.
This is not a new phenomenon. Every major technology has had a phase where the conversation outran the deployment. The difference today is that the gap is being measured in real time, and the measurement comes with a name: the Konstanz AI study.
The Gap Between Hype and Practice
The survey’s findings suggest that the gap between hype and practice is real and measurable. Three percentage points is a small number, but it is also a precise one. It tells us that the rate of AI adoption is not accelerating.
This is a warning sign for companies that assume AI will transform their operations overnight. It is also a signal to workers who believe that the future is arriving faster than it is.
The survey does not predict the future. It describes the present. And the present, according to the numbers, is one where AI is talked about a lot and used by a minority of workers.
What the Survey Does Not Tell Us
The survey has its limits. It is a snapshot of employee behavior, not a comprehensive analysis of organizational strategy. It tells us what workers say they do, not what managers plan or what vendors sell.
It also does not capture the experience of workers who do use AI. We do not know how deeply they integrate it, how often they use it, or whether they find it helpful. The survey only knows that they say they use it.
That is a limitation, but it is also a starting point. Future surveys could ask more detailed questions about how AI fits into daily routines. They could track the same workers over time to see whether the adoption curve changes.
The Bottom Line
The survey’s core finding is simple: AI adoption is not keeping pace with public interest. Workers are curious, but they are not adopting at the rate the headlines suggest.
This is a cautionary note for anyone building a career around AI. The technology is real, but its spread is slower than the stories make it look. The survey shows that the future arrives in dribs and drabs, not in a single dramatic moment.
The gap between hype and practice is real, and it is measurable. The survey measures it precisely: three percentage points over a year. That is the story, and it is the whole story.
Key Facts Box
– Survey: Third wave of the Konstanz AI study
– Respondents: 1,105 employees
– Date: May 2026
– Previous year’s AI use: 35%
– Current year’s AI use: 38%
– Difference: +3 percentage points
Source material: “Workplace AI adoption remains slow, informal and uneven, survey finds,” Phys.org.
Get the Notebook.
The day's best stories and every fresh verdict, in plain English, in your inbox by seven. One email a day, no more.