GoodAI and the AI Roadmap Institute organized a workshop at the ARAYA headquarters in Tokyo on October 13, 2017. The workshop aimed to address the potential pitfalls of a race for transformative AI and foster interdisciplinary discussion on how to avoid them. Workshop participants included individuals from various organizations such as GoodAI, CSER, the Slovak Academy of Science, University of Tokyo, Araya, Peking University, Scientific American, and Dwango AI Laboratory.

The workshop discussed the potential risks of a race for transformative AI, including the disregard for safety procedures and agreements in favor of faster utilization, limited access to the benefits of the technology, and race dynamics regardless of the motivations of the actors. The participants emphasized the need to mitigate these risks by studying the dynamics of the AI race, incentivizing cooperation, building trust through global discussions, avoiding fearmongering, and discussing optimal governance structures for AI development.

To address these challenges, the General AI Challenge Round 2: Race Avoidance will be launched on January 18, 2018. The challenge aims to crowdsource mitigation strategies for the risks associated with the AI race. The workshop and the challenge are part of the efforts to engage more stakeholders and foster cross-disciplinary global discussion on AI development.

The workshop also discussed the progress in AI research and development towards general artificial intelligence (AGI) and the recognition of AI as a strategic priority by various actors. While it is uncertain whether AGI will be achieved in the coming decades, the workshop acknowledged the potential race dynamics and the risks associated with neglecting safety procedures and the unequal distribution of the benefits of transformative AI.

The workshop and the challenge aim to find solutions to the problems associated with the AI race and develop a better understanding of race dynamics. The participants identified the need for frameworks to discuss and formalize AI race questions, the importance of value propagation and resource allocation, the value of clear roadmaps, and the potential use of simulation games to understand the race problem.

The workshop also explored the perspectives on the AI race, recognizing competition as natural and potentially stimulating, and the importance of compromise and consensus. The participants emphasized the need to ensure that the future aligns with most of humanity’s values, while acknowledging the challenges of defining universal human values.

The workshop concluded by discussing the actors involved in the AI race and identifying the need to include more stakeholders in the discussion. The participants highlighted the importance of diverse perspectives and the need to address questions of governance and regulation in AI development.

To get involved with the development of the next round of the General AI Challenge and contribute to addressing the challenges of the AI race, interested individuals and organizations are encouraged to participate and share their ideas and strategies.