Μake us preferred on Google Hundreds of millions of primary and secondary students are starting the new school year this month, and artificial intelligence is reshaping how they learn. Students and teachers alike are asking a basic question: what’s the point of homework when a chatbot can finish it in seconds?
Some countries have moved quickly to bring AI into classrooms. China now requires at least eight hours of AI instruction per year, while the UAE, Saudi Arabia, and Qatar have started weaving AI into their curricula in the Gulf region.
AI chatbots can make schoolwork faster and easier. But the real question is what happens once students set the technology aside, and how much of what looks like learning is actually theirs.
Faster homework, weaker exams
To dig into this, researchers David Stromberg of Stockholm University and Victor Lei and Yanhui Wu of the University of Hong Kong ran one of the largest studies of its kind, following roughly 27,000 students ages 12 to 18 in China over 30 months. Their paper pulled together data on homework, closed-book monthly exams, and high-stakes entrance exams across nine subjects.
The findings reveal a sharp disconnect between completing schoolwork and actually learning from it. After students started using generative AI, their homework scores climbed 18 percent, while the time spent on each assignment dropped by about 30 percent, from an average of 64 minutes down to 45.
But when those same students took monthly exams without AI assistance, their scores fell 20 percent within six months. Their results on high-stakes entrance exams slipped too, with the full impact showing up over a longer stretch of time.
The OECD reached a similar conclusion in its Digital Education Outlook 2026, a review of emerging research published in January. It found that when generative AI isn’t designed for teaching or used within any real structure, offloading tasks to it improves performance without producing any genuine learning gains. Researchers including Dragan Gasevic have described this potential decline in real mental effort as “metacognitive laziness.”
How students actually use AI for school
How students use these tools appears to matter more than whether they use them at all. A survey conducted by Ipsos for the Vodafone Foundation asked 7,000 students ages 12 to 17 across seven European countries how they use AI when studying on their own, without teacher guidance.
The most common use, reported by 56 percent of students, was simply looking up information. Another 45 percent used AI to have concepts or terms explained to them. But 31 percent admitted to using it to get fully completed answers to assignments.
Fewer students used AI in ways that more directly support independent learning: 29 percent used interactive or adaptive content to guide their studying, and only 20 percent used AI to build personalized study plans or track their own progress.
AI’s wider impact on student life
For many students, AI has become more than just a homework tool. It’s increasingly woven into the broader student experience, from getting quick help to simply having something to talk to.
The HEPI/Kortext Student Generative AI Survey 2026, based on responses from 1,054 full-time undergraduates in the UK, found that 49 percent believe AI has improved their overall student experience, citing time saved, better understanding, and instant, personalized support.
The effects go beyond studying, too. For the first time, the survey asked students about loneliness. Most, 59 percent, said AI made no difference either way. Among the rest, opinion was nearly split: 21 percent said they felt less lonely, while 20 percent said they felt more so.
