You are preparing a group presentation: everyone first asks the same AI for an outline, and you end up with four near-identical versions. When two of you disagree, someone types the question into the chatbot. According to a report published on 27 October 2026, this is no longer a lecture-hall anecdote: it is something institutions now have to learn to manage.
What the report says about group work and AI
The Digital Education Council (DEC) and Pearson released Designing AI for Collaborative Learning, a report on how AI can support collaborative learning in higher education. Its starting point is clear: students increasingly consult AI rather than their peers, with fewer peer exchanges, which weakens the mutual interaction collaborative learning depends on.
The DEC release highlights three findings:
- Groups think alike before they meet. Students who prepare with the same AI tools arrive with similar ideas, leaving less genuine disagreement to resolve together.
- AI becomes the group's referee. Many students ask it to settle disputes. They reach a decision without reasoning their way to it and lose the "productive friction" that drives learning.
- Students want guidance. They ask for more guidance on AI in group work, for independent thinking to be protected, and for decisions to stay with the group.
AI that organises, yes; AI that leads, no
The report draws a distinction that matters: students welcome AI that organises and explains, but resist AI that leads the group's thinking. The issue is not the tool itself but the place it takes in the decision.
Another telling detail: some groups set their own rules on AI use, but students doubt their peers would follow them unless assessment required it. The report concludes that institutions need clear policies on AI's role in group work, reinforced through assessment expectations.
"Students are already bringing AI into their group work. Whether it strengthens collaboration depends on how that use is designed."
That sentence comes from Alessandro Di Lullo, CEO of the Digital Education Council, quoted in the release. The report also maps 15 emerging methodologies for using AI in collaborative learning, organised by the stages of collaboration and the role AI plays at each, and examined through seven learning mechanisms.
Why this hits employability skills directly
The release points out that 87% of job postings explicitly ask for human-centric skills such as communication and teamwork (figure from the DEC study Jobs of the Future 2025). Group work is precisely the exercise meant to train them. If AI decides for students, the assignment can succeed on paper without the skill being learned.
We have covered a related angle for Europe: the gap between AI use and AI skills training shows that usage moves faster than learning. The DEC report adds a layer: even when students are trained on the tool, what the tool can erode also needs protecting.
Institutions starting to set rules
The same day, Soongsil University in Seoul announced its "Guidelines for Ethical Use of AI in Education", in effect since 1 October. They sort AI use into three levels: Level 0 (prohibited), Level 1 (limited use) and Level 2 (permitted use). Instructors must tell students in advance what is allowed, including through the syllabus, and the university supplies a syllabus template, an AI use disclosure statement and a checklist. The guidelines also state that academic misconduct cannot be determined solely from AI detection tools, and they include "collaboration" among seven perspectives to build into course design.
The same logic of frameworks over bans appears in France, for example in the University of Rouen's AI charter, which picks rules over a ban. Still, such texts mostly frame individual use. The DEC report points to a less-covered angle: collective use, where AI steps in between the members of a group.
Who wins, who loses
- Possible winners: institutions that write clear rules for group work and tie them to assessment; students who use AI to organise and get explanations without giving it the last word.
- Possible losers: groups where AI is the referee, which may lose the debate and the skill that comes with it; institutions with no framework, leaving each group to improvise rules nobody enforces.
What this means for us, students
These points are our reading, not the report's. For a group project, three simple habits help: confront your ideas before consulting AI, keep the final decision inside the group, and tell your instructor clearly which tool was used for what. The students in the report ask for such rules themselves; no need to wait for them to be imposed. For associations like ours, which run on collective projects, it is also a reminder: a team's value is the disagreement it knows how to resolve.
For another example of what transparency rules change for students, see our article on AI labelling and the AI Act.
Sources: Digital Education Council, "AI Risks Reducing Peer Exchange and Eroding Collaboration, New DEC and Pearson Report Finds" (27 October 2026); Financial News, "Soongsil University to Classify AI Use in Courses into Three Levels" (27 October 2026). Photo: Soongsil University campus, Benlikeshotchocolate, CC BY-SA 4.0 (Wikimedia Commons).
