The Metaverse as an AI-mediated Tool of Targeted Persuasion

Abstract

Over the next five to ten years, the metaverse will transform our digital lives from flat media viewed in the third person to immersive media experienced in the first person. While the core technologies of virtual reality (VR) and augmented reality (AR) are not inherently dangerous to consumers, many advocacy groups and policymakers have raised concerns about the expansive surveillance capabilities that large platforms could deploy across populations. What is often overlooked, however, is how surveillance-related risks could be significantly amplified if metaverse platforms are also allowed to simultaneously deploy targeted promotional content as real-time immersive experiences. When considered in the context of Control Theory (CT), the pairing of real-time surveillance with real-time influence raises serious new concerns, for it could enable interactive platforms to become extremely efficient instruments for targeted deception, manipulation, and persuasion. These risks are further increased by the growing likelihood that AI-controlled avatars (i.e., Virtual Spokespeople) that look, speak, and behave like authentic human users will be used as a primary tool of targeted influence within immersive environments. For these reasons, policymakers must consider limiting the ability of metaverse platforms to deploy real-time immersive influence campaigns, especially when conversational AI technologies are utilized. As described herein, conversational AI paired with real-time behavioral and emotional monitoring could enable profoundly dangerous tools for deploying targeted influence campaigns that compromise the epistemic agency of populations at scale.