The human ability to cooperate in a wide range of contexts is a key ingredient in the success of our species. Problems of cooperation—in which agents seek ways to jointly improve their welfare—are ubiquitous and important. They can be found at every scale, from the daily routines of highway driving, communicating in shared language and work collaborations, to the global challenges of climate change, pandemic preparedness and international trade.
With AI agents playing an ever greater role in our lives, we must endow them with similar abilities. In particular they must understand the behaviors of others, find common ground by which to communicate with them, make credible commitments, and establish institutions which promote cooperative behavior. By construction, the goal of Cooperative AI is interdisciplinary in nature. Therefore, our workshop will bring together scholars from diverse backgrounds including reinforcement learning (and inverse RL), multi-agent systems, human-AI interaction, game theory, mechanism design, social choice, fairness, cognitive science, language learning, and interpretability. Our workshop will include a panel discussion with experts spanning these diverse communities.
This year we will organize the workshop along two axes. First, we will discuss how to incentivize cooperation in AI systems, developing algorithms that can act effectively in general-sum settings, and which encourage others to cooperate. Such systems are crucial for preventing disastrous outcomes (e.g. in traffic), and for achieving joint gains in interaction with other agents, human or machine (e.g. in bargaining problems). In the long run such systems may also provide improved incentive design mechanisms to help humans avoid unfavorable equilibria in real world settings.
The second focus is on how to implement effective coordination, given that cooperation is already incentivized. Even in situations where everyone agrees to cooperate, it is still very difficult to establish and perpetuate the common conventions, language and division of labour necessary to carry out a cooperative act. For example, we may examine zero-shot coordination, in which AI agents need to coordinate with novel partners at test time. This setting is highly relevant to human-AI coordination, and provides a stepping stone for the community towards full Cooperative AI.