Artificial intelligence personalizes learning in Japanese classrooms
Artificial intelligence (AI) is coming to the classroom, bringing tailored instruction for students and a potentially huge efficiency boost for teachers, according to presenters at a Feb. 20 AI and education symposium at Hosei University in Japan’s capital.
Since ChatGPT burst onto the scene in November 2022, it has amassed an estimated 100 million active users and launched frenzied experimentation on how to apply generative AI to numerous fields, education among them.
Kyosuke Takami, a senior researcher at the National Institute for Educational Policy Research, laid out study results involving about 200 high school students that suggest AI could end the days of simply slogging through a textbook and doing hours of exercises before the big test.
The test group was given access to an AI-driven system where they would read explanations of, for example, a mathematical formula and then do exercises to practice, all on a digital device. The AI analyzed their answers and, for the parts they got wrong, generated new explanations and recommend exercises aimed specifically at helping students understand how to get the problems right.
In short, there was no time wasted repeating material already mastered, while the students could also hone in on their weak points. And results showed that the test group students made quicker progress and got better on tests over time than those just doing exercises.
“Instead of just doing a lot of problems, doing those recommended by the AI seems to elevate academic ability generally,” Takami said. He noted, however, that it’s unlikely students and teachers can rely entirely on this system, as some pupils continued to score low on tests even after using it. “Teachers need to get involved to cover the areas AI can’t manage.”
On the teachers’ side of the equation, Hisaho Nakata, a public sector specialist at Microsoft Japan Inc., claimed that generative AIs like ChatGPT can create in-depth lesson and entire course plans tailored to specific learning goals if given a detailed prompt, from structure to images to draft test questions. Not only would this save time, Nakata emphasized, it also has the potential to build students up from mere acquisition of knowledge to synthesis by giving the AI well-considered prompts with that end in mind.
However, when asked about the ethical use of generative AI in classrooms, including the wisdom of introducing the still-new technology without official vetting of its capabilities and outputs, Takami and Nakata demurred, with Takami saying only that discussion was needed.
