Overview

Outline of JSAI

Contents

• Inaugural Address of President
• Address of Editor-in-Chief
• Past presidents
• Board members (JP)
• Representatives (JP)
• Fellow
• Main Activities
• Congress documents (JP)
• Articles of Incorporation of the Japanese Society for Artificial Intelligence
• Information from JSAI
• The Japanese Society for Artificial Intelligence Anti-Harassment Policy

Inaugural Address of President

Toward the Next Decade: The Democratization of AI and JSAI

Yoichi MOTOMURA
(National Institute of Advanced Industrial Science and Technology)

　The “democratization of AI” is accelerating. As used here, the “democratization of AI” means that anyone—not just a handful of experts—can readily use and make the most of artificial intelligence technology, and that the technology is being implemented in the real world and spreading on a broad scale. Now that our Society has reached its 40th anniversary, what does it need in order to move on to the next phase? Artificial intelligence has become commonplace to an unprecedented degree across a wide range of areas spanning academic fields, industry, daily life, and education. The experts and practitioners involved in artificial intelligence have likewise grown incomparably broader and more diverse than ever before. As the democratization of AI advances, interest in a future that makes use of artificial intelligence technology is greater than it has ever been. The number of people involved in artificial intelligence will continue to grow explosively, and their diversity and inclusivity will become increasingly important. It is therefore essential that our Society, as a platform of academic and technical knowledge for research on artificial intelligence, also expand its role and its mission. In other words, our Society itself needs to be “democratized”: everyone, not just a handful of experts, should be able to use it and make the most of it.
　Indeed, our activities directed at a broad audience have grown more vigorous in recent years. The national conference has been expanding in scale year after year, and we have been hosting a great many programs and events. We have also begun distributing videos on YouTube, launched the Junior/Youth membership category, brought our contests together under the “Japan AI Cup,” and begun holding SIAI, our regular industry-academia forum. Our children’s newspaper series, carried monthly by local newspapers throughout Japan in collaboration with Kyodo News, has entered its second year, and the QR codes on the page now take readers straight to explanatory pages on our Society’s website. These activities are the fruit of the ingenuity and enthusiasm of the directors in charge, the committee members, and the administrative staff, and these very qualities are without doubt the great driving force that sustains our Society today. At the same time, if we are to sustain and further develop such efforts, we need to reduce our heavy reliance on particular individuals and systematize our processes so that new participants can join in and collaborate. In other words, if we are to develop our dynamic activities further and make them sustainable, our next major challenge is to build the systems and advance the platform functions that serve as our foundation. This, too, is a democratization of our Society. If we are to preserve what is universal and worth protecting in our Society’s activities while taking on forward-looking challenges, and if we are to keep exploring next-generation ways of running our Society and new initiatives, then building the foundations and the systems that make these efforts easy will be crucial over the next decade. Guided by a forward-looking strategy, we need to imagine our Society as it will be after further growth and put in place the platform that will serve as its foundation.
　With such a strong platform in place, what should the next decade look like for our Society as it marks its 40th anniversary this year? Many of the elementary school children across Japan who have been reading the AI articles in the children’s newspaper will have entered university, and among the students who joined as Junior/Youth members, some will be pursuing full-scale research. The young people who tested their skills against one another at the Japan AI Cup may, through exchanges with the peers and rivals they first met there, be using advanced technologies to push one another to even greater heights. Among the new members who became involved as users of AI technology, some may come to lead development communities by envisioning, from a user’s standpoint, how AI technology should be. Co-creation among researchers and developers who meet at special interest group meetings, specially organized seminars, national conferences that grow larger by the day, and at other similar venues will bring in new methods and technologies from fields beyond the traditional scope of AI and give rise to new forms of AI technology. And with industry collaboration more active than ever, the real-world implementation of AI technology and the transformation of industrial structures will accelerate as well. If we are to set off such chain reactions of emergence, will not diverse and inclusive interaction among members across every generation and every field become ever more important?
　I am convinced that our Society’s role as a foundation and platform that can energize and improve interaction among members is one of the things we cannot do without in the next decade. The signs are already apparent. The directors and committee members believe that interaction among members matters greatly, and they have come to realize the need for an off-site meeting that would enable them to hold long, concentrated discussions. They are currently exploring how to make such a meeting a reality. At the same time, our Director for Information Infrastructure is examining how to make our online platform interactive, as well as how to establish a framework for creating new services and ventures on top of it. Enriching the interaction among our members so that the effect of our growing membership becomes more nonlinear and dramatic will be a major role for our Society as a “place” for interaction. However, a “place” alone does not give rise to interaction. It is the people who gather in a good “place” and carry out their own activities there who are the source of interaction and its true leading players. Our new logo, too, embodies that conviction. Why not become one of those leading players in our Society’s next decade and help create a new future?

Address of Editor-in-Chief

Delivering the “Impossible of a Decade Ago” to Society: The Contribution of Academic Journals in the AI Era

Yoshinobu KANO
(Shizuoka University)

　I am Yoshinobu Kano, and I have been appointed Editor-in-Chief of our Society’s Transactions. AI technology is shaking the very foundations of society, and its reason for existing is being questioned as never before, so to take the helm of the Transactions at such a critical time is a sobering responsibility. First of all, I would like to express my heartfelt gratitude to the members, the reviewers, and the administrative staff who support the Transactions day in and day out. During the terms of my predecessor, Dr. Hirotaka Osawa, Editor-in-Chief of the Transactions, and Dr. Youichiro Miyake, Editor-in-Chief of the Journal, I had the privilege of learning a great deal as an Associate Editor. Together with Dr. Yukino Baba, Editor-in-Chief of the Journal, Dr. Hirohiko Suwa, Associate Editor of the Transactions, and Dr. Sho Yamamoto, Associate Editor of the Journal, I hope to make both the Transactions and the Journal even more valuable.
　Looking back only ten years ago, the rise of generative AI as we know it today and dialogue systems that are almost indistinguishable from human beings were still spoken of as distant dreams. At that time, much of the technology we were painstakingly building kept running up against the limits of accuracy and resources. My own work, with natural language processing at its core, has ranged widely, from the search for human-like intelligence to real-world implementation in public domains such as medicine, the judiciary, and politics. And yet I cannot hide my astonishment at the speed with which the dreams of the past keep turning into reality one after another. To take the research projects I have been involved in myself, with the Todai Robot Project, which asked whether a robot could get into the University of Tokyo by automatically answering university entrance examination questions, the hardest problem of all was getting the machine to read and understand the questions themselves. At COLIEE, the international workshop where systems compete to answer bar examination questions automatically, accuracy is approaching saturation—on the multiple-choice section, at least—to the point where the automation of judicial processes could itself become a demand from society. The Artificial Intelligence-based Werewolf Project, which has machines automatically play the conversational game Werewolf, pursued a goal that at the time everyone regarded as remote, even as a research problem. Today, however, this project stands at the leading edge with teams competing to advance technologies that could translate directly into practical applications. In medicine, the various forms of automated diagnostic support still face many challenges, including scarce data and difficulties peculiar to the domain. Even so, we hear of this technology being put into practical use in one place after another. In addition, we can now see a path toward grasping and predicting people’s opinions and feelings—individually and in aggregate—from their social media posts and speech.
　In short, we are living in a singular age, one in which things that were utterly impossible ten years ago have been realized as part of the everyday landscape. Yet precisely because the technology is growing more sophisticated and more of a black box, the role that a scholarly journal must play weighs heavier than ever. Going beyond only pursuing accuracy, we must articulate and establish the logical explainability of why a given result is produced and the true return to society that arises where diverse fields intersect. One of our Society’s defining features, I believe, is its diversity and its openness. The fields involved in AI keep expanding, and their academic cultures differ enormously. If we are to understand one another’s values across fields, allow them to intersect, and refine them further, what we publish in the Transactions and the Journal matters.
　Our Society covers a wide range of fields, but from the standpoint of engineering, I believe we share a common framework: demonstrating that something is useful through reproducible, quantitative evaluation. At the same time, now that AI technology is coming into wide practical use, the evaluation of research is no longer a matter of competing on accuracy alone. What value does a particular technology return, and what problems can it solve in the complex contexts of the real world? The importance of verifying and describing that many-sided usefulness has likewise never been greater. How are we to connect the fundamental findings born at universities and other research institutions to industry? I want the Transactions to be a platform that accelerates the cycle by which the knowledge gained there is fed back into the deepening of scholarship.
　At the same time, we face a challenge peculiar to our field: because technical progress is so rapid, research results tend to become obsolete in the blink of an eye. Under such conditions, the role the Transactions must play is to be a rapid and dynamic knowledge hub, not merely an archive of information. We must also explore mechanisms that prize timely publication with the capability of sending fresh discussion topics out into the world before they are allowed to stagnate. Fortunately, thanks to the remarkable improvement in translation technology driven by AI itself, the barrier to bringing high-quality research written in Japanese to the English-speaking world has fallen dramatically. Amid an accelerating global race in research and development, I hope that by disseminating insights grounded in Japan’s own contexts and ethics, as well as in meticulous fundamental research, we can build a robust foundation of knowledge that researchers around the world will refer to and that society will trust.

Weaving What Excites Us Now, for the Future

Yukino BABA
(Tokyo University)

　This fiscal year, as both our Society and this Journal reach the 40-year milestone, I have taken up the post of Editor-in-Chief of the Journal, Artificial Intelligence. AI is profoundly reshaping society, and the relationship between AI research and society is being called into question. Under these circumstances, I feel the weight of the responsibility of taking over this important role keenly.
　My first encounter with the Japanese Society for Artificial Intelligence was the 10th Meeting for Youth COMmunity (MYCOM), which I attended in 2009. Since then, I have been involved with our Society in many ways: giving presentations and talks at the national conference, editing this Journal as a student member and later as a regular member of the editorial committee, and helping run our Society as a Director. What I have felt most strongly over these long years of involvement is the freedom of its discussions, its openness to anyone, whatever their field or position, and, above all, the curiosity-driven atmosphere that says, “Let’s do interesting research.” I am delighted that I can now give something back, as Editor-in-Chief, to the Society that has nurtured me since my student days.
　As you know, AI has permeated deeply into both the research setting and society at large. Summarizing papers, explaining techniques, surveying related work—AI can now handle a considerable part of all this. In such an era, what role remains for a bimonthly Journal, in which editorial committee members take time to assemble a special feature, researchers write their manuscripts, and readers pick up the printed page? The body text and equations now in papers will increasingly be organized as data and fed into AI, but the mood of the times that fills the environment where research is actually done—“What are AI researchers interested in right now?” “What discussions are they having?” “Why do they find it interesting?”—is not the sort of thing that survives in the body of a paper. It is precisely the connection of the questions each person was asking and the heat behind them to that particular moment and person, as inscribed throughout these pages, that should ensure that this Journal will remain an irreplaceable primary source when later generations of people and AI look back and ask what kind of place the artificial intelligence research community was in the late 2020s.
　What makes it possible to capture this mood of our own time so vividly is, I believe, that this Journal has placed “freedom of interest” at the heart of its editorial principles. In this Journal, each editorial committee member’s sense of “This is what’s interesting now,” “This is what I want to delve into,” and “This is the person I want to hear from” is reflected candidly on the page. What drew me to our Society when I was a student was precisely that curiosity-driven freedom. During my term as Editor-in-Chief, I want to cherish this freedom and to reflect strongly in these pages the interests of the editorial committee members and, if possible, of our members as well. I look forward to working with you all.

Past presidents