Teens Turn to AI for Homework: Hungarian Study Urges Schools to Teach AI Ethics

A Hungarian study by Eszterházy Károly Catholic University found that 16% of secondary school students used AI tools like ChatGPT for homework, with girls showing more positive attitudes but using AI less than boys. The research highlights an urgent need for ethical AI literacy education to guide responsible use among students.

In a groundbreaking study conducted by researchers Mátyás Turós, Róbert Nagy, and Zoltán Szüts from Eszterházy Károly Catholic University in Hungary, the increasingly controversial role of artificial intelligence in secondary school homework has been subjected to rare, empirical scrutiny. Amid sensationalist media narratives and mounting societal concerns, this research stands out by offering solid data and fresh insights. The survey involved thirty classes across twenty Hungarian high schools, ultimately analyzing 548 essays after a careful data cleansing process. Crucially, instead of relying on widely criticized AI detection tools, the researchers employed human coders, three scholars with PhDs and two PhD students, to detect AI authorship. Their findings reveal that about 16% of Hungarian secondary school students had used AI tools, notably ChatGPT, to assist with or fully complete their homework assignments.

Students’ Sentiments Towards AI: Curious Yet Cautious

The analysis of student essays uncovered a fascinating paradox: while AI is increasingly present in their academic lives, students' feelings about it remain unsettled. Many expressed vague fears about AI being "dangerous" or "sinister," but rarely articulated specific criticisms or thoughtful reservations. Sentiment analysis further confirmed this lack of clarity; attitudes towards AI lacked clear patterns or consistent links to demographic variables. This suggests that among teenagers, AI remains more an object of cultural anxiety than informed critique. Moreover, while many students recognize AI’s potential to transform learning, they also seem unsure how to position themselves ethically and practically in relation to these new technologies. This confusion underscores a profound need for structured AI literacy education, starting from the secondary school level.

Girls Like It, Boys Use It: The Gender Paradox

One of the most striking findings of the study was the gender-based divergence between attitudes and behavior. Female students, on average, expressed more positive views about the role of AI in learning, seeing it as a valuable tool for enhancing educational experiences. Yet paradoxically, they were significantly less likely than their male counterparts to actually use AI to complete homework. This contrast challenges some earlier research suggesting that demographic factors like gender have little influence on AI adoption. It also hints at deeper social dynamics, perhaps linked to risk-taking behaviors, perceptions of academic honesty, or internalized expectations about authenticity and effort. Meanwhile, factors such as age or parents' educational attainment showed no meaningful impact on AI usage, indicating that AI’s integration into education cuts across traditional social boundaries, offering a democratizing potential that schools have yet to fully harness.

How Essays Were Evaluated: Beyond Flawed AI Detectors

Unlike most recent studies that leaned heavily on automated AI detection tools, tools often riddled with false positives and negatives, the Hungarian researchers relied on detailed human coding. Essays were flagged for AI authorship based on several markers, including flawless spelling and grammar, rigid argument structures typical of machine-generated text, and the tendency to justify opinions using bullet points or overly formal phrasing. Human coders also looked for signs of content misalignment with the typical intellectual level expected of 14–18-year-olds. Even with human judgment, distinguishing between human and AI writing was not always straightforward, highlighting the increasingly blurred lines in contemporary academic work. The researchers emphasized that AI-generated texts could easily be tweaked to appear human by introducing minor spelling errors or casual expressions, making detection an evolving challenge for educators.

The Road Ahead: Rethinking Homework and Teaching AI Literacy

The study’s broader implications are profound. The authors argue that education systems must urgently adapt to the reality of widespread AI use among students. They call for integrating critical AI literacy into curricula, teaching not only technical skills but also ethical considerations and the ability to critically assess AI-generated outputs. Schools must set clear institutional guidelines about AI use, and teacher training programs must evolve to prepare educators for a world where AI is an integral part of student learning. The researchers also warn of a looming societal divide between four types of AI users: those who master and grow with the technology, those who passively depend on it, those who reject it, and those excluded from it due to lack of access. To prevent education from reinforcing such inequalities, a balanced and conscious use of AI must become a core 21st-century competency.

Despite its strengths, the study acknowledges certain limitations. Human coding, while more reliable than machine detection, remains imperfect, and the sample, though diverse, cannot claim to be fully representative of Hungary’s entire secondary school population. Nevertheless, this work by Turós, Nagy, and Szüts offers a timely, critical intervention into a rapidly changing educational landscape. It also sets a valuable benchmark for future research, emphasizing the need for longitudinal studies that track not just how students use AI but how it ultimately shapes their cognitive development, ethical sensibilities, and social mobility. As artificial intelligence continues to weave itself into the fabric of everyday life, the lessons from this Hungarian study may soon resonate far beyond the country's classrooms, offering vital guidance to education systems worldwide grappling with the AI revolution.