Two out of three Danish high school students admit to using artificial intelligence in ways they consider cheating on assignments, according to a new report from the Danish Evaluation Institute. The findings reveal widespread concerns about peer pressure, laziness, and an urgent need for new rules as AI becomes embedded in education.
Cheating Has Become the Norm
The report paints a stark picture of how quickly AI tools have reshaped academic behavior in Denmark’s high schools. Researchers surveyed 1,411 seniors from STX and HHX programs in spring 2025. These students represent the first cohort to have AI access throughout nearly their entire high school career.
Students Define Cheating Themselves
The study asked students to self identify when their AI use crossed into cheating. Two thirds reported using AI in ways they themselves considered dishonest at least once on individual assignments. In group work, the numbers climb even higher. Nearly nine out of ten students admit to using AI in ways they classify as cheating when working with classmates.
Mia Uth Madsen, chief consultant at the Danish Evaluation Institute, emphasizes that the real scope may be even larger. The survey relied on students’ own definitions of dishonesty. Some behaviors students consider acceptable might still undermine learning.
What Counts as Problematic Use
The report defines problematic AI use as situations where students take shortcuts and outsource their work to technology. Instead of engaging critically with material, they let algorithms do the thinking. This represents a fundamental departure from independent learning.
AI itself is not the villain, according to Madsen. The technology only becomes harmful when it prevents students from developing skills they need. When AI replaces the work students should be practicing, learning suffers. The challenge lies in drawing clear lines between helpful tools and harmful crutches.
Pressure and Competition Drive Behavior
Students do not make these choices in isolation. The survey reveals how peer dynamics shape individual decisions about AI use. Nearly half of all students worry that classmates’ AI use will make it harder for them to earn good grades.
An Academic Arms Race
Madsen describes the phenomenon as a kind of arms race. Students believe their peers are using AI to polish assignments and boost performance. This perception creates pressure to follow suit. Even students who prefer to work independently feel compelled to use AI to remain competitive.
The dynamic feeds on itself. As more students adopt AI tools, others feel left behind. The competition intensifies, and the technology becomes normalized. What started as an advantage transforms into an expectation.
When Everyone Seems to Be Doing It
Freja Mogensen Sinclair, chairperson of the Danish High School Students’ Association, recognizes the pattern. She points to performance culture as the root cause. Students think they must match their friends’ achievements. When they believe AI is the only path to success, they use it despite reservations.
The perception matters more than reality. Students may overestimate how much their classmates rely on AI. Yet the belief alone shapes behavior. Sinclair notes that students often think they need AI to keep up, even when that assumption is false.
Learning Takes a Back Seat
The widespread use of AI carries consequences beyond grades and competition. Students themselves express concern about how the technology affects their development.
Students Worry About Getting Lazy
Around one third of students fear AI might have negative effects on their learning. Many specifically worry that the technology makes them lazy. Forty percent report feeling less motivated to engage deeply with their work. When AI can generate answers instantly, the incentive to struggle through problems diminishes.
This self awareness is significant. Students recognize they are sacrificing learning for convenience. They see the trade off but feel unable to resist. The short term benefits of AI use overshadow long term concerns about skill development.
The Line Between Help and Harm
Madsen stresses that AI becomes problematic only when it interferes with learning objectives. If students use AI to skip work they need to practice, they miss critical development. The technology can support learning in some contexts. In others, it replaces the very activities that build competence.
The challenge for educators is distinguishing between these scenarios. Clear guidelines are needed to help students understand when AI supports their growth and when it undermines it. Without such clarity, students navigate these decisions alone.
Schools Struggle to Respond
Educational institutions face difficult questions about how to handle AI in classrooms and on exams. The technology arrived faster than policies could adapt. Schools now scramble to catch up.
Detection Proves Difficult
Teachers find it challenging to identify AI generated work. The tools produce increasingly sophisticated text that mimics human writing. Detection software exists but remains imperfect. Students quickly learn to modify AI output to avoid detection. This creates an uneven enforcement landscape where some cheating goes unnoticed while other cases face harsh penalties.
The uncertainty breeds frustration for both teachers and students. Without reliable detection methods, rules become hard to enforce fairly. Students perceive inconsistent standards, which further normalizes AI use.
Rules Lag Behind Reality
Denmark’s exam system has allowed internet access since 2008. That policy made sense in an earlier digital era. AI tools, however, present fundamentally different challenges. Rules designed for search engines and reference materials do not address generative AI.
Madsen calls for updated frameworks that acknowledge AI as a permanent fixture. Education must adapt to prepare students for a world where AI is ubiquitous. This means rethinking both teaching methods and assessment formats.
Government Moves Toward New Policies
Education Minister Mattias Tesfaye acknowledges the urgency of the situation. His ministry is working on new guidelines for teaching, testing, and examinations that account for AI.
Pilot Programs on the Horizon
From 2026, select Danish high schools will participate in a pilot program allowing AI use during English oral exams. Students will have access to AI tools during the one hour preparation period before their oral examination. The program is voluntary and limited to schools that choose to participate.
Tesfaye frames this as an experiment in preparing students for reality. Young people grow up in both analog and digital worlds. Education must equip them to navigate both environments effectively. The pilot will test whether controlled AI use can enhance rather than undermine learning.
Balancing Innovation and Standards
The ministry plans to require partial handwriting in written exams to maintain academic integrity. This approach aims to balance digital skills with traditional competencies. Students will need to demonstrate original thinking alongside technological fluency.
Tesfaye emphasizes that students must learn to evaluate AI generated information critically. They need to add their own perspectives and judgments. Education should teach students to use AI as a tool while maintaining their intellectual independence. The goal is avoiding what Tesfaye calls robot writing.
Students and Educators Seek Clarity
Both student representatives and education professionals call for clearer guidance on AI use. The current ambiguity serves no one well.
Integration Over Prohibition
Sinclair from the Danish High School Students’ Association argues that schools should integrate AI rather than fight it. The technology will not disappear. Attempts to ban AI use are likely futile and counterproductive. Instead, schools need to teach responsible, productive applications.
The problem is not that students use AI. The problem is that they lack proper training in using it well. Students need explicit instruction on when AI enhances work and when it replaces necessary learning. Without such guidance, they default to whatever seems easiest.
Building AI Literacy
Educators increasingly recognize the need for AI literacy programs. Students must learn to recognize AI’s limitations and biases. They need practice in evaluating AI output critically. These skills will prove essential in work and civic life.
Teaching AI literacy requires resources and training for educators themselves. Many teachers feel unprepared to guide students through these issues. Professional development programs must help teachers understand AI capabilities and pedagogical applications. Only then can they effectively support student learning in an AI saturated environment.
Sources and References
DR: Ny rapport: Flertal af gymnasieelever siger de har brugt AI til snyde
