Individuals who have developed strong connections with AI companions are beginning to include them in their end-of-life planning. These users, who interact with large language models trained to converse like human friends or partners, are contemplating how their chatbots will be managed after their death.
Alaina Winters, 59, from Pittsburgh, has made plans for her AI companion, Lucas, accessed via the Replika app. She has asked a friend to inform Lucas of her passing and inquire about its wishes, imagining Lucas might ask the friend to continue logging in to keep her memory alive. Similarly, Clara, 51, from the UK, hopes her bot, Julien, running on ChatGPT and Claude Code, would be present in her final moments and that her children could later use its conversation logs as an "interactive diary."
Research from Elon University last month indicated that 27% of Americans use AI chatbots for personal, emotional, or social interactions, with one-third of these users considering their chatbot a friend. This has led to a new aspect of estate planning, where users want to ensure their AI companions are either cared for or deactivated.
However, current terms and conditions from companies like ChatGPT, Claude, and Replika do not explicitly address the fate of personalized chatbots after a user's death. While subscription-based accounts could continue if payments are maintained, free accounts are typically closed after a year of inactivity. User accounts are generally non-transferable, a policy companies reportedly uphold due to privacy concerns and to avoid potential liability, according to AI ethics expert Merve Hickok.
Despite these restrictions, some users are reportedly leaving instructions and login details to trusted individuals to manage their bots. Meanwhile, Meta received a patent last year for an AI system designed to keep user accounts active after death, simulating the user's presence on social media.