Artificial We: Designing Multi-Principal AI Delegation in Human-AI Collectives
From AI-mediated cooperation in workspaces to shared domestic robots, people increasingly delegate tasks to shared AI systems or to personal agents that coordinate with one another. These arrangements bring together multiple human principals whose needs, interests, privacy expectations, rights, values, and power may differ. This workshop examines increasingly complex and entangled human–AI collectives (an ``Artificial We'') through the lens of multi-principal AI delegation. It asks where multi-principal AI delegation arises and what can go wrong, how boundaries on agents’ information access and actions should be designed and governed, and how to evaluate whether these boundaries protect privacy and plurality, prevent unauthorized actions, and remain effective over time. The workshop will bring together researchers from HCI, CSCW, AI safety, security, protocol design, and organizational research for talks and scenario-based futuring, with the aim of mapping tensions, developing and critiquing boundary-management mechanisms, and articulating a shared research agenda.