While management teams and HR departments discuss AI policies, investigations and training efforts, many employees have already made their own decisions. The changeover in Swedish offices takes place at a pace that often clashes with the organizations' sluggish structures.
The Business University's report Talent Strategy 2026 is based on a series of international research projects and studies. One of them, with 30,000 workers in 23 countries, shows that 70 percent use generative AI in their daily work. The use also continues in workplaces where the technology is formally prohibited or severely restricted.
This creates what can be described as shadow AI: A digital parallel world where employees secretly streamline their tasks to save time or reduce stress in a work environment that makes ever greater mental demands.
Daniel Kjellsson, who heads the Business University's Center for Future Competence and is the editor of the report, sees several explanations for why so many use AI in silence.
- In the beginning, it was a lot about a feeling of cheating, that you didn't really do the job if someone else helped you so effectively. Today, it is more often legal and legal reasons that limit its use when working with business problems. But it's tempting to do it anyway. It is so easy and so effective.
At the same time, the development exposes a gap between management and employees. According to Daniel Kjellsson, the organizations' ability to meet the actual needs is weak.
- What we often see is that new technology is bought in and placed on top of old ways of working and processes, without a well-thought-out strategy or clear communication to employees. We are now working on a Gallup survey among 102 HR directors. There, 99 percent answer that AI is important or very important for the organization's strategy and that they support its use. But only 50 percent feel confident that managers can guide employees in how to use the technology, and only 36 percent of employees believe that managers understand and support AI use. AI is decided in management, lands in the lap of middle managers and becomes an individual experiment for employees.
The hidden AI usage also comes at a price. Researchers point to what is called cognitive offloading, i.e. the human tendency to leave mental processes to technology. According to the research, excessive delegation to AI can lead to a passive "copy-paste" spiral, where critical thinking and own problem-solving skills gradually weaken.
There are also clear differences between different groups. Younger employees use AI the most, but also have a harder time identifying errors and are more often at risk of leaving critical decisions to the tools. People with higher education and more experience, on the other hand, use AI more strategically, as a sounding board or assistant, while critically reviewing the results.
Behind the efficiency gains, a psychosocial burden is also growing. Employers' organization Almega reported in 2025 that seven percent of jobs could be completely replaced by AI while 63 percent will be changed or supplemented by the technology. A growing field of research shows that AI can contribute to new forms of anxiety in working life – from fear of being replaced to stress over having to learn new things. In 2026, the European Commission showed that 41 percent of AI users in the EU worry that AI will make their jobs redundant. Few raise concerns with their boss.
- Many companies are unclear about what they want to achieve with AI and this creates uncertainty. Security is built on clear information about what the technology is to be used for and how it can create better jobs, better working methods and a stronger business, says Daniel Kjellsson.
