The use of AI is increasing – but trust is not following suit

Artificial intelligence has very quickly become part of the private and working lives of many people in Denmark. By October 2023, just under 33 per cent had used generative AI privately. By December 2025, this had increased to almost 62 per cent. Professional use increased from just under 24 per cent to just over 56 per cent during the same period.

This is shown in a report from the Digital Democracy Centre at SDU. According to the head of the centre Professor Claes Holger de Vreese, this development shows that AI has become widespread at an unusually high speed. But greater familiarity with the technology is not necessarily the same as greater understanding of it.

1. What is the most important result of the study?

– First of all, we are seeing a very strong growth in the use of artificial intelligence, both privately and professionally. In a very short time, AI has become part of many different tasks and situations.

– On the other hand, trust in the technology has not increased correspondingly. Although people in Denmark rate their own understanding of AI as higher than before, we cannot conclude from the study that their actual knowledge has increased. There may be a risk that although we feel familiar with the technology because we use it often, we do not necessarily understand how it works or what its limitations are.

2. Why has AI become widespread so quickly?

– The spread of generative AI has been faster than the spread of social media. A significant part of the explanation for this is that the services are easy to use. You can communicate with it in natural language, and you can get started without any training.

– But the easiness can also create a false sense of security. An answer can appear convincing even if it is incorrect or incomplete. At the same time, many AI models are so complex and unfathomable that it can be difficult to explain exactly how they arrived at a particular result.

Some of the groups that use AI the most are also among the most sceptical. The study does not determine whether experience with the technology leads to greater scepticism. One possible explanation is that frequent users encounter the errors and limitations of AI more often.

– Scepticism can be a positive thing but the question is whether it makes us use AI more cautiously or makes us want to learn more. We cannot conclude that from the study.

3. What does this development mean for workplaces?

– In a situation in which the use of AI is growing so rapidly, the responsibility cannot lie only with the individual employee. It is also the role of workplaces to provide their employees with knowledge, guidance and a clear framework for what the technology can be used for and where particular caution is needed.

At the same time, organisations need opportunities to test AI in restricted settings. This could be smaller projects in which employees investigate the possibilities, learn from their mistakes and evaluate the results before an application is used at a larger scale.

– All organisations must learn to deal with AI. This requires not only rules, but also spaces in which employees can use the technology, exchange experiences and learn more about when it creates value and when it does not.

Another study (published in Danish) from the Digital Democracy Centre has shown that Danish people’s support for AI in their interactions with the public sector is related to how often they themselves use the technology and what it is specifically used for. Support for the use of AI is greatest in situations in which AI helps with a limited task, but a human being still makes the final decision.

4. Is there a risk that AI will create new differences?

– The study shows that men, younger people and people with a high level of education are currently among the groups who use AI the most. But the development is happening so quickly that we cannot conclude that these will also be the relevant distinctions in five years.

– It is to be expected that differences will arise with regard to the access to and the benefits gained from AI. We just do not yet know exactly which population groups, professions or sectors will be affected the most. This makes it necessary to continuously investigate who benefits from the technology and who is at risk of being disadvantaged.

5. What should a university take away from the results?

– Universities have a special role because AI affects the entire institution. The technology is ingrained in administrative processes, research and teaching, and the students are already using it. This is why universities have a responsibility for their own use of AI as well as a duty to create knowledge, produce graduates and contribute to the public debate.

This also applies to the impact of AI on the media, information and democracy. The study shows that the concerns of people in Denmark include misinformation and the importance of the technology for their trust in politicians and the news media.

– Discussions about AI are often about innovation, efficiency and new opportunities. But the technology also has an impact on some of the central institutions of democracy. In this respect, universities must generate knowledge, investigate the consequences and help others understand them through research, education and dissemination.

Within the universities, we also need to have an open conversation about how AI is used. In research, the term AI shame is used to describe situations in which people use AI but do not tell anyone about it. This can make it more difficult to share both the good and the bad experiences.

– If the use of AI is kept secret, the organisation will learn less. We must be able to talk openly about where AI helps, where it fails and what professional and ethical questions it raises. There must be clear guidelines but also the flexibility to develop them in line with the technology and our experiences.

For SDU and other universities, the task is therefore to combine critical thinking with the opportunity to test the technology under responsible conditions.

– Recognising that artificial intelligence is here to stay and that it will affect the operations of the entire university is a good place to start. The next step will be to set up processes in which we use the technology, learn from our experiences and continuously consider how it should be included in research, teaching and administration.