Human Digital Twin: Enabling Human-Multi Smart Machines Collaboration

AI solutions are becoming ubiquitous in private and professional Domains. Soon, humans using AI will be relying on recommendations and actions from multiple smart machines to coordinate and manage their financial, professional, or health objectives. As the human's life aspects and objectives are dependent and as her resources are limited, the paper argues that to avoid conflicts, all smart machines supporting a human need to be adaptively aligned with her objectives. The paper introduces the concept of a Human Digital Twin (HDT), which is a human-specific smart machine dedicated to aligning human objectives with the smart machines supporting her. The HDT monitors the entire human-AI space and, based on the human's responses to the various machine actions, the HDT identifies stable human-machine interaction patterns, which can be used to anticipate human responses in given contexts and thus ensure the alignment of the various machines with her objectives. The HDT learns the human-machine interaction patterns by using the structure of the information used in the interactions (metadata), and not by relying on the content of that information. This enables HDTs to overcome the constraint of existing approaches, which require all smart solutions to operate on the same platform. In addition, by relying only on information structures, HDTs would not compromise human data or proprietary provider information.