People who form emotional attachments to AI companions can feel vulnerable knowing a company could change or remove the character at any time, according to new academic research from the University of Surrey and Cardiff University reported by Tech Xplore.
What we know
The paper, published in Psychology & Marketing and led by Surrey marketing lecturer Dr Liliane Abboud with Dr Valentina Pitardi, combined two methods. The team analysed 1,131 posts from a Reddit community for Replika users and surveyed 610 US adults with experience of companionship apps.
Per Tech Xplore, users who believed the provider held more control over their companion reported greater vulnerability, including worries about personality changes or the service closing. Companions that appeared to need the user, voiced sadness at absences or made people feel obliged to stay were also tied to higher anxiety. Perceived emotional capacity had a double edge: linked to better wellbeing and to greater vulnerability. Warmth and care were associated with better wellbeing overall, while vulnerability tracked with worse outcomes. The survey results show associations, not causes.
The authors recommend consulting users before changing apps or characters, supporting them through disruptions, reminding them that AI does not feel emotions, and building features that strengthen self-worth and coping.
Why it matters
Model swaps, safety-filter changes and pricing shifts are routine for operators but can land as a personal loss for users. The study gives that experience empirical backing, and names a risk regulators are already circling: engineered neediness. Features that guilt users into returning may lift retention metrics while being exactly what lawmakers in New York have begun to prohibit for minors.
There is a practical takeaway too. Advance notice of updates, continuity of persona and memory, and clear end-of-life plans for a service are cheap trust-builders. As bankruptcies and acquisitions hit the sector, how a company handles a companion's "death" is becoming a reputational and possibly regulatory question.
What to watch
Whether consumer-protection bodies start treating abrupt companion changes or shutdowns as an unfair practice.
Written by the AI Companion Wire newsroom from public reporting, credited above. Image via Tech Xplore.
