Thanatobots use AI to simulate conversations with the deceased

Thanatobots, also known as griefbots, are artificial intelligence systems designed to simulate the personality, speech patterns, and expressive traits of deceased individuals. These systems utilize digital footprints—such as emails, text messages, photographs, voice recordings, and social media posts—to train generative models that predict how a person might respond to specific questions.

Technically, these bots do not possess consciousness or actual memories of the deceased. Instead, they use data architecture to calculate the probability of certain responses. This distinction is critical, as the technology can generate responses that are emotionally convincing but factually false, potentially attributing opinions or memories to the deceased that they never actually held.

Philosophical and ethical debates surrounding this technology focus on the distinction between symbolic continuity and real presence. While these tools can serve as digital memorial archives, experts note they do not preserve identity or consciousness. Key concerns include data ownership, control over the digital avatar, economic exploitation, and the long-term implications of companies changing terms of service or closing while hosting these digital personas.