Avatarobots and the crisis of democracy: From bot armies to metaverse swarms

In this paper, we argue that immersive virtual reality environments, which we call the metaverse, can reproduce and intensify the risks social media platforms raise for democracies.At the center of our analysis are avatarobots, digital agents that resemble human beings, usually in their speech and behaviour, but sometimes also in their appearance, while operating partly or entirely without continuous direct human control.We examine how AI-powered avatarobots can fabricate synthetic majorities, simulate political agency, and distort the signals on which democratic legitimacy depends.To evaluate these risks, we develop the Principle of Accountable Free Expression (PAFE), structured around three criteria: authenticity, non-domination, and verifiability.These correspond to responsibility for speech, insulation from economic domination, and the verifiable origin of political communication.We use PAFE to analyse the democratic risks raised by avatarobots and to derive design and governance guidelines for immersive environments that aim to function as democratic public spheres.