Who owns my AI twin? Data ownership in a new world of simulated identities
Paulius Jurčys, Mark Fenwick, Ashley Greenwald, Souichirou Kozuka, Valto Loikkanen, Sebastian Porsdam Mann, Brian D. Earp · Computer law & security review · 2026
The advent of AI twins (digital replicas encapsulating individuals’ knowledge, psychological traits, memories, beliefs, preferences, values, and behaviors) introduces unprecedented legal and ethical challenges. These entities, derived from personal data, compel a re-evaluation of what it means to have dominion over one’s data, and even what it is to be a person or to have a personal identity in this AI-mediated age. The argument proceeds in three steps. First, we offer a taxonomy of AI twins and distinguish them from AI assistants and agents. Second, we show that current legal frameworks privilege the technological infrastructure (the “shell”) over the personal data (the “ghost”) that gives each AI twin its distinctive identity and value. Finally, we argue that when individuals consolidate their personal data into self-controlled environments, the classical requirements for recognising property — a clearly defined asset, exclusive control, and value to the owner — are met. Ownership of the underlying data grounds a claim to ownership of the AI twin built from it. As such, the article advocates for a “human-centric” approach to data governance based on a reimagined social contract that prioritizes individual sovereignty and a private-by-default approach to data. This approach seeks to empower individuals, ensure equitable data stewardship, and address the complex socio-legal realities of AI-driven identities.