Scientists, IOs, and multiple stakeholders argue that if and when Artificial General Intelligence (AGI) is developed, it can pose an existential risk to world order, human well-being, and humanity itself. They have proposed multiple solutions to address the problem, such as global governance, international law, and value alignment. However, these and other solutions cannot overcome the difficulty stemming from the search for power by states and the search for profit by high-tech corporations. A nascent global community of practice, including scientists, IOs, civil society, and corporate workers, has been suggesting new practices, beneficial AI principles, an AGI taboo, and how to monitor research on AGI. The advantage of the community of practice over other means of controlling AGI is its capacity to innovate new practices and distribute knowledge and practices across states, the scientific community, corporate practitioners, and global society, thus motivating them to learn why and how to keep AGI inside the bottle.

Emanuel Adler is Professor of Political Science Emeritus, and the Andrea and Charles Bronfman Chair of Israeli Studies Emeritus at the University of Toronto, a Fellow of the Royal Society of Canada and of the European Academy of Sciences, Honorary Professor at the University of Copenhagen, and former editor of International Organization. Previously, he was Professor of International Relations at the Hebrew University of Jerusalem. His publications include books such as The Power of Ideology; Security Communities (with Michael Barnett); Communitarian International Relations; International Practices (with Vincent Pouliot), and World Ordering: A Social Theory of Cognitive Evolution.