The AI Education Paradox: Faster Homework Leads to Weaker Exam Results for Students

Hundreds of millions of primary and secondary students are starting their schools for the new academic year this month, but the advent of Artificial Intelligence is fundamentally altering the learning process. Students and teachers worldwide are grappling with a core dilemma: what is homework for when an AI chatbot can finish it in a matter of seconds?

Governments and educational institutions are responding in varied ways. Some countries have moved fast with the adoption of AI. China requires at least eight hours of AI instruction a year, and in the Gulf, the United Arab Emirates, Saudi Arabia and Qatar have all begun integrating AI into their curricula.

While AI chatbots can make learning faster and easier, critical questions remain about long-term retention. What happens when students put the technology aside? How much of the learning was theirs, and how much stayed with the machine?

To investigate this phenomenon, David Stromberg of Stockholm University, alongside Victor Lei and Yanhui Wu of the University of Hong Kong, conducted one of the largest studies of its kind. They tracked about 27,000 students aged 12 to 18 in China for 30 months, combining data on homework, monthly closed-book exams, and high-stakes entrance exams across nine subjects.

The results revealed a striking gap between doing schoolwork and truly learning from it. After adopting generative AI, students’ homework scores rose by 18 percent, while the time spent on each assignment fell by about 30 percent, dropping from 64 minutes to 45 minutes on average.

However, when those same students sat for monthly exams without AI, their scores fell by 20 percent within six months. Their performance in high-stakes entrance exams also declined, with the full effect emerging over a longer period. The Organisation for Economic Co-operation and Development (OECD) reached a similar conclusion in its Digital Education Outlook 2026, noting that outsourcing tasks to models without proper teaching structures improves performance with no real learning gains. Furthermore, research co-authored by Dragan Gasevic describes this potential erosion of productive cognitive effort as "metacognitive laziness."

How students use these tools proves to be more significant than whether they use them at all. A survey by Ipsos for the Vodafone Foundation asked 7,000 students aged 12 to 17 across seven European countries about their independent study habits. The most common use was getting information, reported by 56 percent of students. Another 45 percent used AI to explain terms and concepts, while 31 percent admitted to using it for complete solutions to tasks. Fewer students utilized AI for independent learning support, with 29 percent using interactive or adaptive content, and only 20 percent using it to create personalised learning plans and track progress.

Beyond academics, AI is increasingly shaping the broader student experience, from instant help to companionship. The HEPI/Kortext Student Generative AI Survey 2026, based on 1,054 full-time undergraduates in the United Kingdom, found that 49 percent believe AI has improved their student experience by saving time, offering better understanding, and providing personalized support. Regarding student loneliness, 59 percent reported no difference, while the remainder split evenly at 21 percent feeling less lonely and 20 percent feeling more lonely.