Classrooms should ban digital textbooks and AI

Oct 18, 2026 - 08:20
 0  9
Classrooms should ban digital textbooks and AI

IN June, Japan's parliament amended the school education law, giving digital textbooks the same legal status as printed books. That means the use of digital textbooks and generative AI will likely increase in schools.

However, the risks of introducing such devices have not been fully discussed, and there are strong concerns that the fundamental linguistic and thinking abilities of students will weaken. If the current situation is left unchecked, we may well lose the talents of coming generations, dealing a huge blow to society.

In 2025, I conducted a survey on "the impact of media choices on writing and reading" in cooperation with the Consortium for Applied Neuroscience, finding an increased shift "away from reading, writing and thinking" among university students in the digital age.

The survey covered 1,062 students, aged 18 to 29, across Japan. Regarding time spent on reading books, newspapers and magazines (including textbooks, manga and official articles on the internet) in everyday life, 20% of respondents said they usually read none of those things at all. In addition, 10% did not take notes during lectures at university or on other occasions, meaning they were students without the habit of writing.

Correct-answer rates on a high school-level reading-comprehension test among non-reading and non-writing students showed that they understood almost none of the content.

Weakened reading skills are possibly caused by a steep decline in the habit of reading long sentences in physical books and taking notes by hand. The superficial acceptance of massive information and video content on the internet and lax understanding of contexts, which result from excessive engagement in short-sentence exchanges on social media, can affect reading abilities.

This trend can apply to adults in general, and is a major risk arising from the digitalized "knowledge economy."

A textbook is an item of teaching material repeatedly read by students to learn a set range of content. Recently, textbooks have included a rapidly growing number of QR codes that give students access to related information sources. However, this has eclipsed the genuine role of textbooks. When students access the internet through QR codes, it becomes difficult to define the scope and depth of learning they need before moving on to the next stage.

It is often cited that a benefit of digital textbooks is that you can easily change the font size to match students' reading preferences. But this feature spoils their memory. Spatial understanding is impaired, making it harder to remember where specific content appears in a book and how pages relate to one another. When trying to position each piece of text in the entire flow and structure of a textbook, physically bound books offer important clues.

A tablet also acts as a brake on learning. The moment you open a physical newspaper, it's likely that articles of potential interest will catch your eye. This perceptual act is hard to reproduce online. The human brain has an excellent function called "parallel search," a cognitive or neural process where the visual system evaluates multiple items simultaneously rather than one by one. That enables finding an article of interest without a keyword search.

Current AI systems do not "generate" something new but merely average existing pieces of data to "synthesise" them. They also tend to create a self-justification bias because they are designed to butter up users by showing pseudo-empathy to prompt them to addictively spend more time on the platform.

There have been tragic reports of people committing suicide after continuously seeking advice from AI. The situation cannot be addressed simply by learning to "go well with AI." As with nuclear power and gene editing, the use of AI requires strict regulation and clear guidelines.

The University of California, Berkeley, School of Law imposed a ban on the use of generative AI for academic work, stressing that independent thinking is essential for educating future lawyers.

From the semester that started in August, it has banned students from using generative AI for most graded dissertation work and examinations. Students are banned from asking an AI tool to draft a paper, brainstorm a paper topic, propose an organizational structure for a paper or summarize a legal rule. Despite the presence of many students from non-English-speaking regions in the master of law course, the use of AI is prohibited for translating a paper originally written in another language into English or for polishing a paper by correcting grammatical mistakes.

Regulation of this kind is needed in all major fields and departments, including science and technology. In Japan, there are countermeasures such as requiring students to handwrite reports and go through face-to-face examinations without the use of electronic devices, but universities have yet to begin deliberations on regulations.

Using AI to write can misrepresent a writer's abilities and betray readers' trust, because the outcome is not truly the author's own and may amount to plagiarism. A lack of awareness of these ethical concerns can foster a false sense of confidence that one could write just as well without AI.

The regulations, which would allow the use of AI only to propose ideas or to correct sentences, do not offer a genuine solution. The more you rely on AI as an excuse, the more you will lose opportunities to improve your abilities and show disrespect for writing and linguistic abilities per se.

A complete ban on AI in education is the only way to eliminate AI's pitfalls and the malignant cycle it can create.

An "AI divide" -- the growing gap in skills and income between those who can use AI effectively and those who can't -- is repeatedly voiced, putting the freedom not to use AI at risk.

In reality, a different divide may be emerging between those whose reliance on AI weakens their ability to think and those who choose not to use AI, preferring to think independently and find their own solutions to problems.

We must bear in mind that any digital or AI technology that restricts the full development of a person's free and independent capacity, such as the creative use of language, is illegitimate, just as it would be to control speech.