아시아출판주간

Asian Publishers Week

Asia Publishing Insight

InterviewJapan

AI Is a Tool. The Power of Story Remains Human.

Toshihiro Sato, who leads AI and data strategy at Kodansha, one of Japan’s largest publishers, explains what the company is building, what it means to protect, and why editors stand to gain.

Toshihiro Sato Director, AI & Data Strategy Department, Kodansha

Toshihiro Sato
Toshihiro Sato

Self-Introduction and the Role of Kodansha’s AI & Data Strategy Department

My name is Toshihiro Sato, and I work in the AI & Data Strategy Department of Kodansha’s IT Strategy Planning Division. After working as a project manager and in other roles in the anime and gaming industries, I joined Kodansha in 2012. I have worked on a range of projects, including developing internal systems and providing hands-on support to business units. I have held my current position since 2026.

The IT Strategy Planning Division is responsible for business development based on digital technology, collaboration with outside companies, the operation and management of internal systems, and infrastructure development.

Within the division, the AI & Data Strategy Department leads efforts to build data infrastructure and develop strategies and governance for the use of AI and data at Kodansha and its group companies.

Specific Examples of AI Use in Publishing

Our work includes AI development projects and sharing information about generative AI.

Development Diary: The RD Nurturing Project

Our department also develops Kodansha’s own AI products. The RD Nurturing Project began with the challenge of exploring whether AI could recreate a character, while we were considering a business collaboration with Tohoku University in Japan. Developed using natural language processing (NLP), it is a project to nurture RD, an artificial intelligence character from the Kaito Queen (Phantom Thief Queen) series, together with fans. We have also received very positive reactions from fans who feel passionately about the character, and the results have exceeded our expectations.

Supporting and Governing AI Use In-House

We also work to raise awareness within the company of AI-related technologies such as those mentioned above.

We have also established a task force that brings together staff from relevant departments to respond to AI-related inquiries and assess the risks of internal AI use. Inquiries come in every day without letup from business, administrative, sales, and other departments, and I feel there is tremendous enthusiasm for using AI within the company.

At the same time, AI services are evolving so quickly that we find it challenging to adapt our internal governance and rules flexibly in response. We have come to realize that introducing AI tools alone is not enough: how thoroughly workflows have been organized beforehand is an extremely important prerequisite.

The Convergence of Roles in AI and Publishing: Future Outlook

No one knows what the future holds, but I feel that, at present, generative AI is by no means producing works that move people in a fundamental way. The ability to create stories that stir readers’ emotions and change their outlook on life remains a uniquely human domain. Genuine achievements by AI in this area are likely still a long way off.

That is why, at a time of rapid technological progress, those of us in publishing need to consider how to engage with AI as a powerful tool and support human creativity, rather than simply whether to let AI create works. There are two sides to this: areas we must protect without compromise, and areas we must boldly seek to transform as times change.

1. Areas to Protect: Creators’ Rights and a Broad View of Risk

Our first priority must be to protect authors’ and creators’ rights and a healthy environment for their creative work. We need to keep pace with questions of legal interpretation, including the flexible copyright exceptions under Article 30-4 of Japan’s Copyright Act for uses not intended for the enjoyment of a work’s expressed thoughts or feelings, as well as the rapid evolution of AI technology, and share what we learn with creators.

It is particularly important to recognize that creators have very different views, feelings, and positions on AI. Some are eager to use it, while others naturally feel anxious or resistant. We urgently need research into how AI can be used safely and effectively in creative work while respecting these diverse views and preserving authors’ rights.

As works are increasingly distributed overseas, we must also pay close attention to differences in audience reactions across countries and regions. Alongside appropriate consideration of national laws, including the EU AI Act, we need to address regional ethical perspectives and cultural reputational risks. Maximizing authors’ rights and interests requires a highly nuanced sense of balance in deciding where to use AI and where to deliberately refrain from using it.

2. Areas to Transform: Digital Transformation and “Agentification” of Publishing Workflows

On the other hand, we need to make bold changes to reduce the workload borne by authors, creators, and their editors. In essence, this means fundamentally transforming publishing workflows through digital technology. Publishing’s long history also means that established practices and labor-intensive manual tasks have often become routine. The key is to establish clear guardrails for acceptable risk while determining how to achieve the “agentification” of these tasks with AI—that is, using AI agents to automate or assist with them. Setting priorities and developing guidelines in this previously uncharted area of AI is an urgent task. The change we seek is an environment in which editors and creators can devote their passion to distinctly human creative work, such as developing new ideas and making a work as compelling as possible, rather than being overwhelmed by administrative tasks.

3. New Experiences and the Evolution of Development Processes

We have also begun researching ways to create new user experiences beyond improvements in operational efficiency. AI-powered conversational bot services, for example, offer fans a richer experience by allowing them to interact with characters rather than simply read the works. We are also beginning to use data and AI, including predictions of a work’s success and assessments of its value, to reach new readers and customers whom we have not reached before, rather than relying solely on past experience. Major changes are also taking place in development. Faster coding with AI has substantially reduced development costs and dramatically shortened the cycle from an idea to a prototype. In response, we would also like to adapt and evolve our existing product management and development operations frameworks to make them more flexible and responsive.

In Closing

No matter how far technology advances, human passion and the joy of creation remain at the heart of a work. By using AI properly and skillfully, we can firmly protect creators’ rights while renewing established systems that need to change, and bring works and readers together around the world. I feel that keeping both of these efforts moving forward is our mission in this era.

In-Depth Follow-Up Questions and Answers

Question 1. When considering potential risks associated with internal AI use, which assessment criteria do you prioritize most highly?

We place the greatest importance on two aspects: legal and technical considerations. On the legal side, we closely monitor the latest case law and the implementation of laws in each country, and rigorously assess the effects on copyright holders’ rights and the publishing industry. On the technical side, we assess technical stability and security risks, focusing on major updates to leading LLMs and similar developments.

Question 2. As you move toward the “agentification” of publishing workflows with AI, could you share one specific task or framework that you are currently prioritizing for practical application?

The fundamental prerequisite is to establish strong governance covering the prevention of internal data leaks, information security, and copyright clearance. We, too, are only now reaching the stage of beginning full-scale work on designing specific assessment criteria and operating rules.

Building on that strong governance, our highest priority for workflow digital transformation is an “Expanded Search System for Published Works.” In a publishing business that handles a wide range of titles, well-timed and precisely targeted sales initiatives directly affect revenue. With so many works entrusted to us every day, we want to use AI to improve the accuracy of information about their release dates, core themes, and detailed target audiences from multiple angles, so that they reach readers rather than becoming buried in the market.

For example, adding accurate metadata about works to semantic search systems or the knowledge bases searched by AI agents can fundamentally prevent appealing books from becoming buried in a vast catalog and enable us to reach the readers they should reach.

At the same time, editors will be freed from administrative tasks such as organizing data and entering attributes, allowing them to concentrate on developing substantive sales strategies. We also expect significant benefits for creators, as their works gain exposure in the most suitable form and their sales opportunities are maximized.

Note: These are my personal views and do not represent the views of Kodansha.

About the Contributors

Toshihiro Sato Director of the AI & Data Strategy Department in Kodansha’s IT Strategy Planning Division.

Asia Publishing InsightAsian Publishers Week 2026

Published by
Korea Publishers Society (KOPUS)
Supported by
Ministry of Culture, Sports and Tourism; Korea Publication Industry Promotion Agency
Website
asianpubweek.org

Copyright of the texts in this webzine belongs to the individual authors and KOPUS.

Scroll to Top