Artificial Intelligence Has Become Normal in Higher Education, but Universities Are Falling Behind
Artificial intelligence has already become a familiar part of education, but universities have still not managed to rebuild their rules, assessment systems and overall approach to teaching. Students actively use AI to prepare assignments, search for information and analyze material, while many higher education institutions are only beginning to shape policies for its use. This was shown by the international AI in Higher Education Global Survey 2026, published by the Digital Education Council. The report was presented on July 9. It involved 45,398 respondents from 35 countries, including 27,284 students and 18,114 educators. The study became one of the largest global surveys on the use of artificial intelligence in higher education. Its results demonstrate the gap between students’ actual practices and the speed at which universities are able to change curricula, academic integrity rules and methods of assessing knowledge.
Time for Action analyzed how AI is changing education, what risks students and educators already see, and why universities need not formal bans, but clear rules for working with new tools. According to the survey, 88% of students already use artificial intelligence tools in their studies. Among educators, the figure stands at 77%, which is 16 percentage points higher than a year earlier. These figures show that AI is no longer an experimental tool for isolated groups. It has become part of the everyday learning process. The problem is that university rules and teaching methods are changing much more slowly. More than half of students, 57%, said they do not receive sufficiently clear guidance on the use of AI when completing assignments. Only 29% believe that their educators are well prepared to explain how to work properly with such tools. At the same time, 64% of educators reported that they had already completed AI literacy training. This indicates that some preparation is taking place, but its results have not yet become visible to most students.
“Artificial intelligence entered the daily lives of students and educators faster than universities were able to adapt. Widespread use has already become the norm, but there is still no coherent practice for its implementation,” the report’s authors note.
AI provides students with real advantages. Some 61% of respondents said that these tools allow them to spend more time analyzing and understanding ideas. Another 31% have started completing more complex academic assignments. This means that artificial intelligence can be more than a way to obtain an answer faster. It can also serve as a tool for working with large volumes of information, identifying connections and testing hypotheses. However, its benefit depends on how exactly the student uses the technology. The study also recorded the first signs of excessive dependence. Some 22% of students admitted that it is becoming more difficult for them to work without artificial intelligence. Another 20% rely on it to complete assignments to a higher standard. A further 19% noticed that they had become worse at remembering educational material. These figures do not mean that AI automatically worsens learning. They point to another problem: when a tool begins to replace independent work rather than support it, students gradually lose some of the skills they are supposed to develop.
Critical thinking is a particular concern. Some 66% of students fear that the use of AI may make learning more superficial. Among educators, more than 73% share this view. The risk is that a ready-made answer can create a sense of understanding without deep engagement with the material. A student receives a structured text or solution but does not always go through the process necessary to form an independent position, verify arguments and identify mistakes.
Another problem concerns academic integrity. Some 60% of students worry that their peers may use AI to gain an unfair advantage. This issue is more complex than ordinary cheating. The boundary between permitted assistance and the replacement of independent work often remains unclear. One educator may allow AI to be used for generating ideas, another only for language editing, while a third may completely prohibit any such tools. As a result, students may be studying under different rules even within the same university. The greatest delay can be seen in assessment systems. Only 28% of students believe that university assignments help them develop the skills they will need for future work with artificial intelligence. Another 24% reported that they are not allowed to use AI in any assessed assignment. This approach may temporarily simplify oversight, but it does not prepare students for a labor market where such tools are already widely used. As a result, 37% of students are not confident that their education program corresponds to the future demands of the professional environment. Among educators, 43% do not believe that their courses may become outdated by the time students graduate. This difference in perception shows that students feel the pressure of labor market change more strongly. They already see that employers expect them to know how to work with AI, analyze its responses, verify data and use technology without losing their own professional competence. The report’s authors also refer to the results of another Digital Education Council study. In 2025, 80% of employers said that higher education was failing to keep pace with changes taking place in industry. This means that the problem extends beyond individual universities. If curricula are not updated, students may receive a degree without gaining practical skills that have already become basic requirements for modern work. The study also identified regional differences. In countries across the Asia-Pacific region, Europe, the Middle East, Africa and Latin America, most educators plan to continue actively using AI in teaching.
In the United States and Canada, the share of educators planning to continue actively using AI fell from 76% to 67%. This was the lowest figure among all regions included in the study. However, these data should not be interpreted as a complete rejection of the technology. They are more likely to indicate a more cautious attitude toward its impact on the quality of education, academic integrity and students’ independence. In the United States and Canada, 55% of educators believe that artificial intelligence may pose a threat to human intellectual development. In addition, 43% of students would support a complete ban on the use of AI at universities. Such attitudes may be related to the fact that universities in this region respond more sharply to the risks of automating thought, replacing authorship and making knowledge assessment more difficult. At the same time, a ban alone does not solve the problem if students are already widely using AI outside formal rules. The main conclusion of the study is that universities need a systematic policy. Students must understand when AI may be used, for which assignments it is acceptable and where the boundary lies between assistance and a breach of academic integrity.
Educators, in turn, need more than short AI literacy courses. They require ready-to-use teaching methods, updated assignment formats and clear assessment rules. Traditional written assignments that can easily be completed using generative AI are becoming less effective at showing a student’s actual knowledge. Universities will therefore have to make wider use of oral defenses, practical assignments, project work, source analysis and explanations of the reasoning behind students’ own decisions.
“Students are not asking for lower standards they are asking for clear rules. Educators want to adapt, but they need an institutional vision. The universities that begin acting now will shape the future of education instead of chasing change,” Alessandro Di Lullo and Danny Bielik conclude.
Artificial intelligence has already changed the way students search for information, complete assignments and prepare for classes. The main task for universities now is not to try to return education to the state it was in before generative AI appeared, but to build rules that allow the technology to be used without reducing the quality of learning. How quickly universities update their programs, assessment systems and academic integrity standards will determine more than the quality of education. It will also determine whether graduates are able to work with new technologies professionally, critically and responsibly.
Sources: Digital Education Council, EdTech Innovation Hub.
About the Author
Related posts:
Artificial Intelligence and Science: Has AI Come Closer to Nobel-Level Discoveries? 
Artificial Intelligence in Ukraine: Why It Is Not a Must-Have Today but Will Shape Careers Tomorrow 
How Artificial Intelligence Discovered Over 12,000 Potential Antibiotics in Ancient Microbes 
Europe Is Looking for Its Own Path in AI: Why Artificial Intelligence Has Become a Matter of Sovereignty, Security and the Economy
