Konstantinos Komaitis and Rumman Chowdhury argue that autonomous AI agents are forcing a rethink of internet governance, shifting the central question from what the internet does to who is acting through it. They contend that emerging interoperability standards, including Anthropic’s Model Context Protocol and Google’s Agent2Agent protocol, need to converge around principles of verifiable human delegation to preserve accountability. The authors call for systems in which delegation stays visible, permissions are meaningful, and every agent action traces back to an authorizing human.
