A study conducted at Germany’s IU International University found that, in a single month in 2025, 57.6% of its 76,485 distance-learning students used Syntea - an assistant embedded in the university’s learning platform and powered by GPT models. More than 44,000 people turned to it to clarify questions, engage with course materials, prepare for exams or receive tailored guidance.
Uneven adoption
Generation Z recorded the highest level of use: 63.66% used the assistant during the period examined. The figure was 51.39% among millennials and 50.27% among Generation X students.
Another factor: younger students are more likely to be in the early stages of degree programmes, where Syntea is more widely available. Those working on dissertations, by contrast, are often unable to use it.
Available when needed
Syntea is used mainly on weekdays and during daytime hours, with activity peaking at around 11 a.m. Part-time students, however, show relatively greater use in the late afternoon, evenings and at weekends.
What does this mean? An AI assistant can extend academic support when lecturers, tutorials and other university services are unavailable. Its potential lies in adapting to people who must combine study with work, family responsibilities and limited time.
A complement, not a substitute
Round-the-clock availability does not turn AI into a replacement for teaching. The study found relatively lower take-up in Architecture and Construction, and in Design and Media—fields in which learning relies more heavily on visual practice, projects, professional tools and applied work.
For now, conversational assistants appear better suited to text-based, conceptual and structured material.
The implication: universities should avoid imposing a single technological solution on disciplines and learning practices that are fundamentally different.
The gender nuance
Usage was marginally higher among women: 59.05%, compared with 54.94% among men. This differs from earlier survey-based research, in which men reported greater affinity with chatbots.
But caution is needed. The mix of degree programmes, the availability of the tool across modules and the way it is integrated into the curriculum may explain as much as—if not more than—individual attitudes, the authors note.
What to bear in mind
Activity records do not establish that students learn more effectively, become more independent, remain enrolled for longer or achieve higher grades. Nor do they assess the quality of responses, factual mistakes, possible over-reliance or the ways students verify AI-generated information.
And while the research focuses on one online university, one month and a platform developed by the institution itself, its scale places it among the largest studies of AI learning assistants in higher education.
Its central contribution: it shows that these systems can already become part of the everyday routine of a large university community.
This article is republished from Futuribles. Here's the original article in Spanish and English.