Writer Profile

Tadahiro Taniguchi
Professor, College of Information Science and Engineering, Ritsumeikan University; Representative Director, Bibliobattle Association
Tadahiro Taniguchi
Professor, College of Information Science and Engineering, Ritsumeikan University; Representative Director, Bibliobattle Association
Bibliobattle is a book introduction communication game that anyone can easily participate in. I first proposed it in 2007, and in 2010, the Bibliobattle Promotion Committee was organized by volunteers. Since then, various enthusiasts have continued to emerge, and it has spread throughout Japan. It carries the catchphrase: "Know the book through the person. Know the person through the book." Currently, the "National University Bibliobattle," a national competition for university students supervised by the Yomiuri Shimbun, has continued for over 10 years (with a hiatus during the COVID-19 pandemic). However, the main focus of Bibliobattle is not such large competitions, but rather the small communication spaces held in neighborhood cafes and school classrooms.
The rules of Bibliobattle are simple. They are defined as follows:
1. Presenters gather with a book they have read and found interesting.
2. Each person introduces their book in turn for five minutes.
3. After each presentation, all participants engage in a discussion about that presentation for two to three minutes.
4. After all presentations are finished, all participants cast one vote each based on the criteria of "Which book did you want to read the most?" The book that receives the most votes becomes the "Champ Book."
Anything that does not follow these rules is not called a "Bibliobattle." A game is defined by its rules. It is the same as saying that soccer where you are allowed to use your hands is no longer soccer.
Bibliobattle is sometimes perceived as a method for promoting reading, but essentially, it is a method for "creating a place to encounter books" and a mechanism for "knowledge sharing." Since there are not enough words to discuss this in detail here, I would appreciate it if you could refer to my books such as "Bibliobattle: A Book Review Game to Know Books and People" (Bungeishunju), "Kamogawa Communication School: From Bibliobattle to Artificial Intelligence" (Sekai Shisosha), and "Mechanism Design of Communication Fields" (Keio University Press).
Now, times are changing. Since the release of ChatGPT at the end of 2022, a massive boom in generative AI has been occurring. This is also affecting the field of reading promotion, which is closely related to Bibliobattle. Book reports, which have often been compared to Bibliobattle, are being directly hit by this influence. There are stories like "a student had ChatGPT write their book report."
The Japan School Library Association, which organizes book report contests for elementary, junior high, and high school students, quickly announced a ban on generative AI in March 2023. "Book reports," which involve reading and understanding the content of a book, summarizing it or writing impressions in written language, and submitting it to an unseen recipient, are actually something that ChatGPT excels at.
What about Bibliobattle? Book reports and Bibliobattle are different. In Bibliobattle, the text generated by the presenter is not the goal. Its essence lies in the dynamics of the communication space where people recommend books, read them, empathize with each other, and get to know one another. Therefore, Bibliobattle currently does not prohibit the use of generative AI, including ChatGPT.
The voting to decide the Champ Book in a Bibliobattle is determined by the subjectivity of everyone present regarding whether they "wanted to read it." It is meaningless and will not resonate unless it is the presenter's own words spoken after having read the book themselves. A speaker who uses words generated by AI to express thoughts they do not actually hold will lose the trust of those around them.
About 10 years ago, I conducted research to clarify the nature of Bibliobattle as a "book recommendation system" by "performing comparative experiments with information recommendation methods such as content-based recommendation and collaborative filtering." Through this research, I reaffirmed the fundamental difference between existing information recommendation systems and Bibliobattle. While many information recommendation systems refer to past reading or purchase history to guess "You'll like this, right?", in Bibliobattle, the recommendation is made according to the presenter's—the "person's"—own "love," saying "I like this. It's interesting, so I'm sure you'll like it too."
Generative AI is still nothing more than a tool. It is convenient and excellent, but it cannot become a human subject. Generative AI can generate reports from book information. Generative AI has destroyed the institution of the book report, which existed on that same dimension. However, Bibliobattle is a mechanism for a communication space that creates the dynamics of interaction where people share knowledge, and therefore it does not exist on a dimension that can be destroyed by generative AI.
Generative AI is intensely forcing us to reconsider our intelligence and what it means to be human. Even in this era, Bibliobattle will continue to be a mechanism for a communication space where people "know books through people and know people through books." Fifteen years have passed since its inception, but its true value is surely yet to be tested.
*Affiliations and titles are as of the time this magazine was published.