AI in classrooms debate
Schools and ed‑tech are wrestling with how to stop AI from giving students answers they don’t understand, and a new platform aims to address that exact problem by preserving learning rather than just producing outputs. A survey in Japan found 73.7% of high‑school students already use conversational AI—mostly as a study assistant—underscoring how quickly the issue has moved from novelty to classroom routine. (it-online.co.za) (english.kyodonews.net)
Students are already using artificial intelligence for schoolwork at scale, and schools are shifting from trying to block it to trying to make it teach. (bernama.com) A survey by the Gakken Research Institute for Learning and Education found 73.7% of Japanese high school students use conversational artificial intelligence, along with 43.2% of junior high students and 36.6% of elementary school students. The online survey, conducted in November 2025, drew responses from 2,400 students across grades one through 12. (bernama.com) Among Japanese high school students who use the tools, 42.3% said they use them to help with studying and homework, and 26.0% said they use them to find information. More high school students also said their thinking ability had become weaker rather than stronger after using generative artificial intelligence. (bernama.com) That is the classroom problem schools are now trying to solve: a chatbot can produce an answer in seconds, but a student can still miss the reasoning that gets there. UNESCO said in its 2023 guidance that education systems were largely unprepared to validate these tools and called for a human-centered, age-appropriate approach. (unesco.org) Japan’s Ministry of Education, Culture, Sports, Science and Technology updated its school guidance in December 2024 with Version 2.0 of its generative artificial intelligence guidelines. The ministry said schools should treat chatbot output as a reference, keep human judgment at the center, and use the tools only when they support curriculum goals and students’ information literacy. (mext.go.jp) The debate has also widened beyond copying and cheating into questions about privacy, age, and assessment. UNESCO’s guidance called for data-protection rules and even an age limit for unsupervised conversations with generative artificial intelligence platforms. (unesco.org) Schools that rely on detection software face their own limits. Turnitin said its document-level false positive rate is under 1% only for documents with at least 20% artificial intelligence writing, while its sentence-level false positive rate is about 4%, meaning some human-written sentences can still be flagged. (turnitin.ca) That is why the newest education technology pitch is not just “spot the bot” or “generate the essay.” The selling point is a system that slows students down, shows steps, and keeps the learning process visible enough for teachers to judge whether understanding is actually happening. (news.harvard.edu) Common Sense Media reported in September 2024 that young people were already adopting generative artificial intelligence at home and at school, often without much parent awareness or teacher guidance. Two school years later, the Japanese survey suggests that for many students, the chatbot is no longer a novelty but part of routine study. (commonsensemedia.org, bernama.com)
