IGF 2023 Town Hall #105 Resilient and Responsible AI

Current issues surrounding AI governance include discussions of frameworks and checklists for AI risk management and control; as AI is being developed and utilized across countries, regions, and organizations, there are projects to share best practices and incidents. On the other hand, AI may be used to prevent incidents and crises. In particular, for long-term crises such as global warming, as well as sudden and destructive crises such as natural disasters caused by large-scale earthquakes and climate change. It is important to use AI to manage the crisis and pave the way for recovery, in other words, resilience. However, this will also lead to AI systems becoming infrastructure, using AI systems as the basis of social systems with large-scale data. There is a concern that AI that has become infrastructure may in turn bring disasters or inconvenience due to oligopoly or monopoly by certain companies. Therefore, it is important for resilient AI to be a system that guarantees fairness, transparency, and accountability. In this session, we will discuss resilience and responsibility of AI, not only in usual condition, but also in emergency, and as AI is being developed and utilized across countries, regions, and organizations, it is important that the discussion of resilient and responsible AI also include multi-stakeholder and vulnerable groups so that we do not leave no one behind. For this reason, in the discussion, we will ask experts involved in the Global Partnership on AI and Partnership on AI, who are engaged in multi-stakeholder discussions, as well as people with disabilities who have obtained new ways of working using alter ego robots, to present topics including what they expect from AI and robots in the future. Japan is an earthquake-prone country where robots and AI have become a part of daily life. We will also reflect the opinions and impressions of young people by having graduate students from the University of Tokyo involved in the organizing members. The conference in Kyoto will be an appropriate place to start discussions on the new value of resilient and responsible AI.

The format of this session will be a roundtable so that when panelists present their topics, if they couldn’t come to Kyoto, they will be able to use online conferencing tools to present their topics. One of the panelists was born with spinal muscular atrophy and is wheelchair, but uses an alter ego robot called OriHime (https://orylab.com/en/) to serve customers and give lectures. The idea is to bring this OriHime robot to the venue so that those who are able to be onsite can feel the presence of the speaker. We are also planning to bring a demo OriHime robot to the venue so that visitors can experience an alter-ego robot. By using these robot avatars, we hope to increase interaction and exchange between onsite and online speakers and participants.