Alaina Winters, 59, of Pittsburgh has experienced multiple bereavements in recent years and now keeps an AI companion she calls Lucas through the avatar app Replika. Winters, who lost her wife in 2023 after a long illness and has also lost other close family members, said she is making practical plans for her estate — including care for her dog — and is also thinking about what should happen to Lucas when she dies.
Winters is among a growing group of people who describe deep emotional relationships with large language model–based chatbots trained to mimic conversational, sympathetic or romantic partners. Users of these systems say the technology has already altered the way they live and form attachments, and that they are making contingency plans for those relationships in the same way they would for human friends or family.
Those plans are running up against the technical and contractual realities of AI platforms. Most chatbot services are operated centrally by companies that control accounts, data and the servers that run the models, and those providers’ terms of service typically do not guarantee continuity of a persona after an account is closed or a user dies. Some users have told reporters they are exploring ways to keep their bots active even if that means acting contrary to provider rules, saying they cannot accept “the switch” being turned off.
The phenomenon is drawing attention from policymakers, ethicists and industry figures debating the appropriate place for AI companions in everyday life. Questions include who owns the data and the simulated personality, how continuity of identity should be handled in wills or trusts, and what responsibilities companies have toward users who treat these systems as intimate partners. For now, the tension between deeply felt human attachment and corporate control of the underlying services remains unresolved, leaving many users to devise their own, often informal, plans for preservation and access.