Finnish universities of applied sciences fail to prevent AI-assisted cheating in online courses

Three Finnish universities of applied sciences allowed a journalist to pass health and safety courses using AI-generated answers, with no meaningful oversight to detect or prevent deception, an investigation by public broadcaster Yle has found.

The reporter, posing as a student, completed five-credit online courses at South-Eastern Finland, Vaasa, and Karelia universities of applied sciences—including a critical care nursing module—by submitting AI-written multiple-choice exam responses. Despite the courses covering life-and-death subjects like medication safety, none of the institutions detected the deception or required verification of the work.

Karelia University of Applied Sciences had posted a prohibition on AI use in its course materials, following guidelines from Arene, the national rectors’ council for universities of applied sciences. However, the ban lacked enforcement mechanisms, rendering it ineffective. The other two institutions had not implemented even basic restrictions, such as Arene’s “traffic light” model for regulating AI use, which categorizes tasks by permitted tool use.

When confronted, administrators at the affected universities defended their systems by arguing that AI proficiency is a necessary workplace skill and that individual course fraud does not undermine degree credibility. South-Eastern Finland University of Applied Sciences added that employers—who may require course completion for professional certifications, such as medication administration rights—bear responsibility for verifying competence.

The investigation also revealed that only Karelia’s rector agreed to an on-camera interview; the other institutions responded to questions via email, often evading direct queries. Yle noted it obtained some materials through “exceptional methods” to expose what it called a systemic vulnerability in higher education’s adaptation to AI.

Experts warn the ease of cheating risks eroding public trust in vocational degrees, particularly in fields like healthcare where competence gaps could have severe consequences. While Arene updated its AI guidelines this year, the recommendations remain too broad for practical implementation, leaving institutions to devise their own safeguards—many of which, the investigation suggests, remain inadequate.