New government guidelines on the use of AI in education
Students must comprehend the characteristics of AI before they use it, according to the guidelines which explicitly state that passing off reports, essays or other works produced by AI as one’s own is inappropriate.
The government guidelines, which are being described as ‘provisional’, are expected to set-off a flurry of activity using AI in higher education starting with increasing AI literacy among teachers. The guidelines are also aimed at reducing workloads by streamlining administrative tasks and improving teaching practices.
Education Minister Keiko Nagaoka, during a 4 July news conference announcing the new guidelines, said the government was committed to the use of generative AI, including automated writing systems such as ChatGPT, that helps promote English language learning and encourages integration into group activities, enhancing diversity learning.
“Based on the guidelines that protect (students and teachers) from the disadvantages (of generative AI), we are committed to addressing concerns, enhancing teacher understanding and skills, and fostering a safe and effective environment for AI utilisation in education,” Nagaoka said.
Japan’s education globalisation drive is focusing on expanding English language capacity and encouraging experience among students in a country that is still struggling to increase international exposure.
Teachers are also expected to discard traditional exams and reports that can be easily done using generative AI, according to Hisanobu Muto, a school digitalisation project team leader at the Education Ministry.
Universities seeking clarification
Universities have been seeking clarification, with different views being expressed at home and abroad about the speedy adoption of generative AI. Some countries ban the use of ChatGPT by students, and others allow it, with some controls. In Japan different universities have adopted different policies.
A major concern among educators in Japan is that students will not be able to acquire the ability to think or express themselves if generative AI is used inappropriately, such as for writing essays.
In a statement in May the Japan Association of National Universities urged universities to strengthen its use while controlling negative aspects. The Association also cited concerns about overdependence on generative AI in writing academic papers, calling for each university to establish clear rules on its use.
For example, in April the University of Tokyo posted a document on its internal website on Generative AI, which also referred to other interactive AI, stating “reports must be created by students themselves and cannot be created solely with the help of AI”.
Tohoku University also posted a notice on using generative AI on its website, and urged faculty to rethink assignments and how exams are administered. It recommended “checking how AI will respond before assigning exercises and reports” and “switching to an exam format that requires students to write in the classroom”.
Tohoku, in instructions to teaching staff in March, said it was not realistic to completely ban the use of generative AI but allowed staff some leeway, saying: “It is desirable that teachers themselves think about how to approach generative AI, such as whether to partially ban it or to provide instructions to students on how to use it.”
The education ministry has said the guidelines will be updated based on emerging experiences and options, allowing for some flexibility.
Need to balance advantages against risks
The government has set up an AI Strategy Council involving university professors, legal experts and businesses involved in AI, chaired by Professor Yutaka Matsuo, a leading AI researcher at the University of Tokyo, to discuss the policies needed for AI. It held its first meeting at the prime minister’s office in May, which included several cabinet ministers. The Council characterised AI-related changes as the “arrival of a great opportunity” for Japan.
However, Matsuo also cited seven risks of generative AI, including the issues of sophisticated crime, copyright infringement, a flood of false information, and inappropriate use of personal information, among others.
The Council is prioritising drawing up newly formulated regulations and legislation. Prime Minister Fumio Kishida wants the strategy to be supported by the government budget, giving it more scope for speedy implementation.
In an editorial published on 3 June, Yomiuri Shimbun, Japan’s largest circulation daily newspaper, cautioned the government against promoting the use of AI too quickly. It pointed out that Japan has fallen behind other countries in digitisation and wanted the government to “restore the nation’s position” with regard to AI “but it is wrong to rush headlong into utilising AI while leaving consideration of the risks on the back burner”.
Surveys show that students are already using ChatGPT
However, students are already adopting generative AI, regardless of the pace of regulation.
Japanese media last month reported the results of a Tohoku University-led survey of 4,000 undergraduate students, which found that approximately a third have used ChatGPT. Around 91% of them reported relying on ChatGPT to check and correct their answers in assignments, while more than 85% relied on AI for editing and generating text and to add sentences to express their own ideas.
The results “help highlight the weak points in Japan’s higher education, such as less focus on developing writing skills”, said Fujio Omori, an education policy expert at Tohoku University, who led the team from different universities that conducted the survey over ten days ending on 2 June. Overall, the students were positive about using ChatGPT for their academic work.
Another survey of 19 Japanese universities conducted by the newspaper Mainichi on 22 April, pointed to universities already preparing guidelines for students that include prohibiting students from using AI-generated text, programming code and calculations results in reports, essays and dissertations.
However, most universities did not state an explicit view on the use of generative AI, saying they were gathering information with formulating guidelines in mind.