Half of Japanese University Students Have Used Generative AI; 30% Use on Regular Basis

Yomiuri Shimbun file photo
A student uses generative AI.

One in two university students has used text generating artificial intelligence such as ChatGPT, and 30% of students use such AI on a regular basis, according to a survey of about 10,000 students conducted by the National Federation of University Co-operative Associations (NFUCA).

ChatGPT, created by the U.S. company OpenAI, was released free to the public in November 2022 and has seen adoption worldwide. In Japan, use is growing among university students, the survey found.

A total of 9,873 undergraduate students at 31 national, public and private universities in Japan responded to the survey conducted from October to November. This was 59th survey in a series and asked about the use of generative AI for the first time.

Students who have used ChatGPT or other generative AI accounted for 46.7% of respondents, with 28.9% using such services frequently and 17.8% not currently using them. Students who have never used generative AI accounted for 50.6% of respondents, with 28.2% saying they wanted to use such services in the future and 22.4% saying they wouldn’t use them.

Asked what they used the services for, with multiple answers allowed, 22.1% of students said “as references for writing papers and reports,”12.1% said “for translation or essays in foreign languages,” and 11% said “for consultations and chats.”

A number of Japanese universities have banned unauthorized use of generative AI in papers and reports since last spring.

“Going forward, the specific ways in which people use [generative AI] could spread by word of mouth, leading use to spread rapidly,” said NFUCA’s executive director. “Teachers are currently in a trial-and-error process as they instruct their students.”

Reports copied from AI

“The number of students using AI is increasing,” a professor at a private university in the Kansai region said.

AI-generated reports do not include data or viewpoints mentioned in class, and this makes them immediately recognizable. Even so, student after student is submitting reports copied entirely from AI replies, without any sense of compunction.

The professor requires students to write based on what they have learned in class, but said, “With the emergence of AI, we have to change the very nature of the exam itself.”

Since generative appeared, a number of universities have warned their students not to submit AI-drafted reports. Last July, the Education, Culture, Sports, Science and Technology Ministry urged universities to establish new rules, asserting that having AI write reports “would not deepen students’ learning” and “could be considered plagiarism.” However, students are not fully aware of the dangers inherent in AI.

A 22-year-old senior at a private university in Tokyo uses AI texts to write reports by rearranging the sentence order or replacing formal expressions with plain language.

“It feels the same as searching the internet,” she said. “I don’t feel guilty about it. Once you get used to the convenience, there is no going back.”

On the other hand, a 22-year-old junior at another private university in Tokyo worries that such casual use may foster disregard for originality and copyrights.

“It’s dangerous to just have technology evolving while laws and morals fail to keep pace,” he said.

Motohisa Kaneko, a professor at the University of Tsukuba, said, “It will become even more important for teachers to devise questions that cannot be answered by AI alone. At the same time, they also need to teach students about the dangers of AI, such as the possibility of false information being included in the generated text.”

The Yomiuri Shimbun