GPAI Future of Work: Insights from International Survey

Event Overview

　Established in June 2020, the Global Partnership on AI (GPAI) is an international initiative dedicated to the responsible development and use of artificial intelligence (AI) based on a “human-centric” view. One of the working groups discusses the “Future of Work.”
　As a project of this working group, we are conducting international interview surveys across the world on how our work styles will change as AI is introduced into the workplace. This survey is unique in that the interviews are conducted by students.
　The event invited Dr. Yuko Harayama, former 2020-2021 co-chair of the GPAI’s “The Future of Work” and Dr. Yann Ferguson (Toulouse Institute of Technology, France), who is leading the interview survey, as topic speakers. Dr. Arisa Ema will also give an overview of the survey conducted in Japan this fiscal year, and together with designated discussant Prof. Hideaki Shiroyama, a panelist will discuss the characteristics of the GPAI survey and its prospects.
 

About GPAI – Dr. Yuko Harayama

　Global Partnership on AI (GPAI) is an international framework that aims to achieve the development and use of “responsible AI” based on a human-centered approach and principles such as transparency and respect for human rights. GPAI includes worldwide industry, civil society, governments, and science and technology experts. Multi-stakeholder discussions will be held, and activities will be undertaken to connect theory and practice regarding AI and its future.
Initially discussed at the G7, the summit “Enabling the responsible adoption of AI” hosted by Canada in 2018 and the summit “Technology for Humanity Conference” hosted by the French Presidency in 2019 “Tech for Humanity Meeting” led to the idea of launching a new international forum called GPAI. In May 2020, the G7 Science and Technology Ministers’ Meeting (held online in the U.S.) agreed to cooperate in its launch, and the GPAI was established in June of the same year.
　The GPAI has four working groups, and today I would like to introduce one of them, “The Future of Work.” I myself was co-chair until last year, and now Matthias Peissner and Uday Desai are co-chairs. The “Future of Work” working group aims to deepen our understanding of the impact of AI on our work and to consolidate our findings. To this end, we are focusing on examining practical cases and exploring the concept of a living lab for a vision of the future. We will then investigate the specific impact of AI on workers and the working environment. Of the several projects within the working group, today, we will be discussing use cases.

 

Observation platform in the workplace – Dr. Yann Ferguson

　We launched this platform to understand the impact of AI on the workplace and workers, and conversely, the impact of the workplace and workers on AI. We are cataloging use cases for AI systems by interviewing people in various categories, including executives and developers, or people in the public, private, and nonprofit sectors and startups in diverse industries.
　When we began our research in 2020, we used a mixed approach of online surveys and interviews. However, in 2021, we systematized the interview survey, as interviews provide deeper insights. The interviews consist of five pillars: (1) motivations for implementing AI systems, (2) people’s involvement in defining AI systems and in the design and development process, (3) the role of Human-Machine Interaction in AI system implementation, (4) ethical considerations, and (5) the impact of AI systems on employment, work, and organizations. The GPAI survey is unique in that the student community, rather than consultants, will conduct the interviews.
　The results of the survey showed that most of the AI systems surveyed were in the proof of concept (PoC) stage. Furthermore, even if they were working well in the PoC stage, it did not translate into productivity. Actually, AI systems have to face three challenges to be permanently deployed: reorganizations, as AI reorganizes organizational activities; socialization, as AI destabilizes value systems related to activities; and practice, as AI transforms or destroys professional activities during professional practices. Conversely, PoC is important in helping people imagine what AI is like and understanding the characteristics and potential of AI systems. Moreover, it creates organizational learning effects.
　Based on these findings, we made three recommendations for the 2021 report. First, establishing methodological principles for the PoC of AI systems; second, empowering workers to utilize AI, including training them; and finally, considering fair AI by reducing data bias and establishing independent ethics committees.
　Future developments include a digital prototype of this observation platform in collaboration with the living lab project, and then disseminating this observation platform for more complete case selection and classification. We will also work with student communities in Japan, Brazil, India, New Zealand, and other countries, as well as conduct further research and analysis of cultural specificities in the use of AI in the workplace.
 

The overview of the “Future of Work” Japan Survey – Dr. Arisa Ema

　I would like to present an overview of the “Future of Work” survey conducted in Japan. Before that, I would like to briefly outline the industrial structure surrounding AI in Japan.
A country’s industrial structure is an important factor when considering AI and work. Japan is characterized by numerous business-to-business (B2B) companies rather than business-to-consumer (B2C) companies, which are closer to AI service users. Therefore, AI development, data acquisition, and service provision may all be performed by different companies. Then, for example, when an accident or incident related to AI occurs, it is difficult to determine who will take responsibility in the long supply chain and how far back to intervene in the supply chain to prevent the problem from recurring. In Japan, this industrial structure hinders the introduction of AI services.
　Additionally, Japan has a remarkably high percentage of information technology (IT) personnel belonging to IT companies, exceeding 70%. Hence, there are few IT professionals outside of IT companies, which is among the reasons why it is difficult for companies to introduce AI into the workplace.
　Despite this situation in Japan, there are several examples of AI already being introduced into the workplace. Nine students, including undergraduate and graduate students, participated in the survey in the fall and winter of 2021. While the students were the main actors in the survey, we organized a management team including university researchers to provide support. This was necessary to present the GPAI to the students and to build trust with the companies we interviewed.
　Although the interview items were the same as those used in the international survey, we added our own questions according to the interests of the interviewees and students. For example, these included the social issues behind the use of AI and changes in people’s awareness due to the use of AI.
　The companies we interviewed differed widely in fields and organizational forms. Because of the short time frame of this year’s survey, we positioned it as a preliminary survey and asked for interviews with companies that the students were interested in, rather than covering all industrial fields. Consequently, we received responses from the following industries: finance, public administration, infrastructure and construction, translation, telecommunications and broadcasting, and nursing care.
　The results show that many companies cited labor shortages due to an aging workforce, changes in work styles, and social changes as reasons for introducing AI. In addition to technical issues such as safety and accuracy, they also reported social issues such as fairness and privacy. Compared to the survey conducted by Dr. Ferguson and his team, many of the interviews in Japan were conducted with companies that are actually operating services, rather than in the PoC stage.
　The GPAI survey is characterized by its student-driven approach, and feedback from the students was largely positive regarding their participation in an international project.
 

Panel Discussion

　The panel was asked questions by the designated discussant, Prof. Hideaki Shiroyama of the University of Tokyo. The first question regarded the organizational aspects of the GPAI, which comprises initiatives and governance structures by institutions such as the Organisation for Economic Co-operation and Development (OECD) and G7, but is led by experts at the working group level. Dr. Harayama was asked about the characteristics she identifies in such a hybrid organization.
　Dr. Harayama, who has managed GPAI as co-chair, said that the GPAI organization resembles building a start-up company. She was given full discretion regarding what to discuss, and her first task as co-chair consisted in conducting discussions among the members to identify the main issues to be addressed. Regarding how the GPAI will be institutionalized, it was decided that it would be an international organization, with the OECD acting as the secretariat for high-level meetings and the Centers of Expertise in Paris and Montreal acting as the secretariat for working groups. Dr. Harayama pointed out that it will be extremely interesting to see whether it will shift to an international organizational style or maintain a start-up mindset.
　Prof. Shiroyama’s next question concerned how to assess the future impact of technology. Students are the main actors in the GPAI survey, and he wondered if any new insights can be drawn from their involvement. He also asked how to connect the observation platform with the living lab.
　In response, Dr. Ferguson highlighted that the interviewees want the students to understand and create a positive relationship, just as seniors help juniors. By linking the observation platform to the living lab, we are also trying to create a place where visitors to the living lab in a virtual space can understand and analyze the social, technological, and organizational impact of AI on their work. In other words, he answered that the living lab itself aims to become a place where interviews can be conducted and feedback can be obtained.
　As Dr. Harayama pointed out, GPAI currently resembles a start-up company, and it will require many ideas and participation to connect the platform to the living lab. Therefore, the event ended with the message that GPAI is looking forward to your participation!

Arisa Ema (upper row left), Hiroaki Shiroyama (upper row right), Yuko Harayama (lower row left), Yann Ferguson (lower low right)