For the third consecutive year, Croatian citizens participated in a national survey on the perception of artificial intelligence (AI), and the results clearly show – awareness of artificial intelligence is growing, but trust in its application still seriously lags behind. The research was conducted by a team of scientists led by Assoc. Prof. Dr. Robert Kopal, in collaboration with Asst. Prof. Dr. Krešimir Žnidar and Dr. Darija Korkut, and was realized by the company Prizma CPI in September 2025, on a representative sample of 2,500 adult citizens. The organizer and sponsor of the research is Algebra Bernays University, where the presentation of the mentioned research also took place. Under the moderation of Robert Kopal, the panel included Andrea Čović Vidović, Deputy Head of the European Commission Representation in Croatia, Ratko Mutavdžić, Head of HUP’s Coordination for Artificial Intelligence, Zdravko Kunić, Head of the University Department of Data Science and Artificial Intelligence at Algebra Bernays University, and Marijana Bačić, Member of the Management Board and Chief Operating Officer for Business Customers at Hrvatski Telekom and Director of Combis. Almost all respondents (95 percent) have been exposed to news and information related to artificial intelligence in the past six months. AI literacy, measured by a special scale, shows a slight increase compared to previous years. Citizens demonstrate greater awareness and more frequent use of AI tools, but at the same time express concerns about their use. When it comes to the first associations with the term “artificial intelligence,” negative reactions (54 percent) still outweigh positive ones (40 percent). However, the trend shows a slight decline in negative attitudes and an increase in positive ones, while the share of undecided remains high – even 42 percent.
According to Mutavdžić, the fact that many who use AI do not actually understand it is concerning. He also raised the issue of individual digital literacy. On the other hand, Kunić commented that it is a good sign that there is now more discussion about AI.
– It is natural to fear things we do not understand, but we must not allow AI to shape us; we must shape AI – he said.
From ChatGPT to Everyday Assistance
More than half of citizens (54 percent) use AI tools at least occasionally, which is a significant jump compared to 2024 (35 percent). The most commonly used are ChatGPT, Perplexity, and Copilot, and the application is expanding from the business to the private sphere. AI is increasingly used as a substitute for traditional search engines and as a tool for quick access to specific information. Despite this, most users stick to free versions, while only a smaller portion of citizens considers paid options. Interestingly, the number of those who use AI exclusively privately is growing, indicating a shift in the perception of artificial intelligence – from a professional tool to an everyday companion.
– We do not talk enough about the opportunities and obstacles regarding AI. AI is a good tool that can help in hiring, selection, and data analysis, although the final decision should always be made by a human, not artificial intelligence – concludes Bačić.
Benefits and Risks
Particularly concerning is the fact that an increasing number of citizens believe that the risks of artificial intelligence outweigh its benefits (33 percent compared to 22 percent). This marks a complete turnaround compared to the first two waves of research in 2023 and 2024, when citizens were more optimistic. The greatest benefits continue to be highlighted as the ability to replace humans in dangerous jobs, while the most pronounced risks are manipulation and malicious use of technology.
Trust and Fears
Trust in AI systems stagnates: only 21 percent of citizens trust their application, while 37 percent express distrust. Interestingly, Croatian AI solutions instill less trust than foreign ones, at a ratio of 2:1.
Citizens’ fears are focused on three areas: disinformation, privacy, and the future of work. As many as 59 percent of citizens fear that they will not be able to distinguish AI-generated content from that created by humans. Nearly three-quarters fear that AI will encourage the spread of fake news, and 46 percent expect significant job losses in the next five to ten years.
– I would liken this time to some past eras when certain technologies appeared that were foreign to people at first, and later became ubiquitous. People think of different things when they hear the word “AI.” My mom says it’s a robot, my colleagues at work say it’s a “voicebot.” There is a big difference in how people view AI, and with that come differences between private and business users. The EC wants to simplify regulations around AI – said Andrea Čović Vidović.
AI in Education and the Future
The education system is simultaneously viewed as a space of opportunities and threats. AI is seen as a help to students, but there are concerns that it could negatively impact creativity, mental health, and the quality of knowledge. A particular challenge is the disinterest of citizens in additional AI education. Namely, 36 percent show no interest at all, while 30 percent express willingness. The low level of awareness about state programs further complicates the situation – only a quarter of citizens have heard of vouchers for digital and green skills, and half of them only know that the program exists. The third year of research clearly shows: artificial intelligence has become an integral part of the lives of Croatian citizens, but trust and understanding still do not keep pace with the level of use. Croatian society faces the challenge of how to leverage the advantages of technology while simultaneously protecting citizens from risks. Experts emphasize that education and the development of digital skills must become a priority to bridge the gap between technological capabilities and societal readiness to accept them.
– We must strengthen both the right and left leg; it is a problem if we only strengthen one. Why do we not have a subject “critical thinking,” why do we not have a subject “creativity”? And you will see this a little later, among employers in the top three places, and we cannot do that with technology. When I looked at Algebra’s syllabus and the entire curriculum for artificial intelligence with Zdravko, it is designed exactly to give you both the right and left leg – concluded Kopal at the end of the panel.
