People, Protocols and AI: Power and Governance in Open Social Networks

The window for meaningful coordination is narrow. Agent protocols are evolving rapidly, and the architectural choices made in the next 12 to 18 months will be difficult to reverse. The open social network community has, despite still considerable challenges, built up precisely the institutional knowledge needed about what goes wrong when identity management is poorly designed, when consent is treated as an afterthought, and when governance is added retrospectively. This knowledge is essential for designers of agentic AI protocols, and this workshop is designed to put it to optimal use.

The workshop builds on the report "People, Protocols and AI: Power and Governance in Open Social Networks" (Knodel, Nicole, Sigal, Tan & Theodule) from March 2026, and the workshop held in New Delhi, organised by the Observer Research Foundation during the AI Impact Summit.