Dr Johnny Chan, a senior lecturer in information systems at the University of Auckland Business School, has published an opinion piece arguing that AI companions strip out the disagreement and unavailability that give human relationships their meaning. The commentary, dated 16 October on the university's news site, was first published by Newsroom.
What we know
Chan's central target is the common defence that a companion bot is "better than nothing" for lonely or grieving people. He argues that there is rarely truly nothing after a loss, pointing to memories, photos, messages and the shared history held by other people.
He gives particular attention to griefbots and digital replicas, citing the case of Sunshine Henle, who in 2023 used ChatGPT and old text messages to build a chatbot version of her late mother. Chan suggests a distinction between fictional companions and AI copies of identifiable people, which raise questions of consent and a person's right not to take part. He also asks what happens when short-term comfort turns into ongoing reliance.
The piece is an argument rather than research: it cites no surveys or data and proposes no specific regulatory fix.
Why it matters
The critique goes to the heart of companion product design. Many apps are tuned for agreeableness and constant availability because those traits drive retention. Commentary like Chan's, alongside wider concern about sycophantic chatbots, gives regulators a vocabulary for treating that design choice as a potential harm rather than a feature.
His distinction between fictional characters and replicas of real people is the most practical takeaway. Products that let users recreate a deceased relative or an ex-partner sit in a distinct risk category, touching consent, likeness rights and vulnerable users at once. Operators offering such features should expect closer scrutiny and may want explicit consent and opt-out mechanisms before lawmakers require them. Building in some deliberate friction, such as boundaries, breaks or encouragement towards offline contact, could become a defensible differentiator.
What to watch
Whether griefbot and real-person replica features become a specific focus for legislators already writing companion chatbot rules.
Written by the AI Companion Wire newsroom from public reporting, credited above. Image via University of Auckland.
