“It’s not a question of pitting man against machine, but of finding the ideal collaborative relationship between the two”
Dr. Yuko Harayama is Director of International Affairs at Riken, and co-chair of the Future of Work project at GPAI
What is the mission of your working group?
Our mission as a working group of 34 experts and one OECD observer was to work on the impact of artificial intelligence in the world of work, based not only on the productivity metric, but also on other criteria, such as the working conditions of employees. Our intention was focused on existing practices before moving on to consider the contours of what artificial intelligence might look like in the workplace of the future. Consistent with the GPAI project as a whole, our ambition is to close the gap between practice and theory.
What were the main priorities for your work?
This year, our research is focused on exploring the role of man/machine interactions. It’s obviously not a question of pitting one against the other, but of finding the ideal collaborative relationship between the two. We have also explored the issue of bias in artificial intelligence, because the ways in which data are selected and algorithms are written are not always neutral. Using artificial intelligence to improve working conditions requires data that are as close to reality as possible. We also felt that training was a crucial factor in the way AI is seen in the workplace: it’s important that employees have the tools they need to help them come to terms effectively with this new reality, and maintain their skill levels.
The development of an observation platform is one of your key projects. So how would you describe the current status of that project?
This platform has been built from a number of case studies gathered through the corporate connections of our experts. Drawing on the expertise of one of our sociologists, we were able to prepare questionnaires which we then e-mailed to companies, and followed up in parallel with face-to-face interviews. One of our goals was to analyse the impact of specific cultural characteristics on the implementation of AI in companies, based on case studies conducted in different countries with different cultures.
What conclusions were you able to draw from this project, and what are your recommendations?
Our preliminary analyses suggest that the implementation of processes driven by artificial intelligence is still in its infancy. In some companies, the results of doing so have not lived up to expectations. In others, the success achieved with initial experiments has yet to be confirmed in the longer term. This is because the introduction of AI demands a complete redesign of company working methods, which in turn requires a relatively long adaptation and socialisation period for employees to come to terms with these methods. As part of our considerations, we have also assumed the creation of ethics committees outside these companies to provide objective opinions about their artificial intelligence practices.