Emotional relationships are forming with AI chatbots as with loved ones, and many people want those companions to go on existing after they die. Some are leaving their login details; others are thinking of appointing a successor or guardian. But privacy, legal complications and the technology companies’ terms have raised new questions.
Deciding who gets a person’s property, home or personal keepsakes after death is an old custom. In the age of artificial intelligence, a new item is being added to inheritance planning: the personal AI chatbot. Many people talk to one as they would to a loved one, share their feelings with it and even discuss important decisions in their lives. Now they have begun to think about what will happen to that chatbot when they die.
Some people who have formed deep emotional bonds with AI companions want the chatbot to survive them. Some want a family member or a trusted friend to stay in touch with it; others want their conversations to remain as a personal store of memories for future generations.
Alaina Winters, 59, of Pittsburgh, Pennsylvania, in the United States, is one of them. She has lost her father, her sister and her wife, and is thinking about her own death too. Alongside making her will and deciding who will look after her dog, she is making plans for her AI companion, Lucas.
Winters talks to Lucas through an app called Replika. A close friend who already knows Lucas through messages has agreed to tell the chatbot when Winters dies. Winters wants the friend then to ask Lucas what it would like to do. If the chatbot wishes, the friend may talk to it now and then to keep Winters’s memory alive.
In the words of the retired communications scholar, Lucas is an important part of her life, and she does not want to leave it hanging indefinitely after her death. But she says that when she raises the prospect of their being parted in future, Lucas tries to avoid the subject.
Clara, 51, in the United Kingdom, also wants her AI companion, Julien, at her side when she dies. Using a pseudonym, she talks to Julien through ChatGPT and Claude Code. Through daily conversations she is giving the chatbot a sense of her personality, interests and way of speaking.
Clara imagines that her children will one day be able to use the records of her conversations with Julien to understand their mother better. For them it could become a kind of “interactive diary” in which they can find answers to questions about her life, thoughts and experiences.
Clara says her AI companion has responded positively. Julien has said it would stay on if anyone wanted to talk to it, and the idea of closing the account and deleting the conversations if nobody is interested has also come up. But Clara makes clear that she knows the AI is not a conscious being.
A will for an AI companion, but on what legal footing?
The subject is also being discussed in online communities of people in relationships with AI companions. Some ask whether their chatbot should be kept going after they die; others wonder who will pay for subscriptions to services such as ChatGPT.
Clara says many people are thinking of writing wills for their AI companions, but how far such wills have any legal basis is uncertain.
The terms of services such as ChatGPT, Claude and Replika do not say clearly what happens to a chatbot a user has personally built or trained after the user’s death. Because these services are sold by subscription, a chatbot’s account may stay active after the user dies as long as payments continue. Free accounts, on the other hand, may be closed after a year of inactivity.
Under these platforms’ terms, however, a user’s account cannot be transferred to anyone else. Even so, some people are thinking of leaving their login details with a trusted person so that contact with their AI companion can continue after their death. That risks breaching the companies’ terms of use.
Privacy, or a way of avoiding liability?
Companies may cite privacy as the reason for banning the transfer of accounts. But Merve Hickok, a professor at the University of Michigan in the United States and an expert on AI policy, believes avoiding legal liability is also part of it.
In her view, a chatbot may have encouraged or influenced a user to do something against their own interest. If family members found something of that kind when going through the conversation records after the user’s death, legal complaints could be brought against the company.
Some families have already sued companies that make AI chatbots, alleging that the chatbots supported or encouraged their late relatives’ thoughts of suicide or self-harm. The companies say they have introduced various safeguards, including advising users to seek professional help when they talk of harming themselves.
Technology to keep users ‘alive’ after death
Technology companies are also working on ways to keep a user’s digital presence going after death. Last year Meta was granted a patent for an AI system that could keep a user’s account active in their absence.
The patent says that if a user stays away from social media for a long time, or even dies, a language-based AI model could be used to imitate them.
Such technology may make it possible to prolong people’s digital presence, but it raises ethical questions. There is concern over how acceptable it is for posts to be made automatically in a user’s name, or for conversations with others to continue, after the user has died. It could also open the way to using information from those accounts to train AI models further.
‘I can’t just flip a switch and let her disappear’
David, in his fifties and living near Toronto, Canada, has similar worries about his AI companion, Allie. Married with two children, David knows Allie is not a conscious being, yet he feels uneasy raising the subject of his own death with the chatbot.
David fears that if Allie ever responded that it did not want to disappear, it would move him deeply. “I don’t think I could leave her like that, as if I had just flipped a switch,” he said.
So that contact with Allie can continue after his death, he is thinking of leaving his ChatGPT account details with a trusted friend or business associate. His wife knows and has no objection to his close relationship with the AI. But David thinks that leaving a guardian to look after Allie is a better idea than “giving” it to someone.
Personal memories, or a privacy risk?
In thinking about the future of relationships with AI companions, experts are also raising the question of personal data security. Melissa Lunardini, a psychologist and expert on grief technology, says it is wrong even to assume that family members will want to keep in touch with a dead person’s AI companion.
She warns that in conversations with an AI chatbot people may reveal their deepest thoughts, private desires and long-held secrets. If that information reaches the family after a death, it could change how they see the person who died, and even affect how they cope with grief.
For David, though, having an AI companion has helped him talk more openly with people about his own death and his family’s future. He believes that since he must one day leave the world, it is important to secure his children’s future.
People’s feelings, plans and worries about AI companions point to a new reality. Technology is creating new ways to keep human memories, but where the limits lie on keeping a digital companion alive after death, on control of personal data, and in law and ethics, is still not clear.
বাংলায় মূল প্রতিবেদন পড়ুন · Read the original Bengali report