AI and personal data in the metaverse: a risk assessment

Artificial Intelligence, in combination with other Metaverse technologies such as AR/VR/MR, can create realistic virtual worlds and engaging experiences. Body worn sensors allow human movements, reactions, and even emotions to be recognised by AI models. This enables a user’s behaviour in the real world to be relayed to their avatar in the Metaverse automatically. The collection and processing of such personal data is a cornerstone of Metaverse development. However, security, privacy, and safety are common concerns amongst users, and protections must be built into the Metaverse’s architecture to ensure that individuals and their privacy are kept safe. Policies, regulations, and procedures are needed to enact user protection, privacy, and security. Current academic literature on AI use in the Metaverse reviews various AI techniques and maps them to opportunities, but the associated risks and threats are not analysed. Conversely, surveys on the use of biometric data within AI systems are not specific to the Metaverse and the rich sources of information associated with cyber-physical environments. This paper aims to deepen the conversations around personal data in the Metaverse and expose the risks associated with AI processing.