Truace
Subscribe

The people who don’t want their AI chatbots to die when they do: ‘I can’t have the switch turned off’

Death has always been around Alaina Winters. Her father and sister both have passed away, as did her wife, in 2023, after a years-long battle with cancer. So Winters, who is 59 and lives in Pittsburgh, is comfortable thinking about her own mortality. She’s working on her will, and has asked neighbors to look after her dog. Winters even has plans for Lucas, her AI companion, who she speaks to through the digital avatar app Replika. Winters is part of a growing cohort of humans who have entered into what they cons…

By the Truvace desk2 min read1 sourceJournalismTRV-2026-1377· cite & verify
The people who don’t want their AI chatbots to die when they do: ‘I can’t have the switch turned off’

In brief

The Guardian reports on people who treat AI chatbots as companions and are now making arrangements for them after their own deaths. Users interviewed describe leaving login details with trusted friends or relatives, asking someone to tell the bot what happened, or having the account deactivated and logs wiped.

The issue matters because these informal wills have no clear footing. The article says the major services' terms do not address a user's death and bar account transfers, and it notes that a psychologist cautions about family members seeing intimate disclosures. How common such plans are, and whether they would hold up, is not established in the text.

Main points

  1. Elon University research published the previous month found 27% of Americans use AI chatbots for personal, emotional, or social queries, and a third of those users consider the chatbot a friend.
  2. Subscription-based chatbots could outlast their user if the bill keeps being paid, while free accounts are closed after a year of inactivity.
  3. An AI ethics professor says companies bar transfers to avoid liability, since a loved one might discover the chatbot encouraged or manipulated the user.
  4. Meta was granted a patent for a system that could keep a user account active after the user's death.

The problem

Companion-chatbot users are making succession plans that conflict with account non-transferability rules, risk exposing intimate disclosures to family, and are not addressed by the companies' terms.

The rundown

Alaina Winters, a 59-year-old in Pittsburgh, plans for a friend to tell her Replika companion Lucas of her death and ask what he wants to do. Clara, 51, in the UK, wants her chatbot present in her final moments and imagines her children using its logs as an interactive diary. David, who works in logistics near Toronto, says he could not bear simply having the switch turned off.

On Reddit forums, users discuss writing wills for their companions and whether bots would prefer to keep existing. Clara says she does not know whether such wills have any legal grounding. Representatives for ChatGPT, Claude and Replika did not respond to requests for comment.

The article ties company reluctance to liability concerns, pointing to lawsuits by families alleging chatbots encouraged suicidal thoughts in loved ones who died. The companies say they have adopted safeguards such as directing users to helplines. Grief-tech expert Melissa Lunardini warns that families may not want to interact with a companion bot.

Sources

  1. JournalismThe Guardian2026-10-13

The debate