When an AI Companion Remembers a Child

In an essay published on 19 November 2026, psychologist and psychoanalyst Tom Wooldridge asks a difficult question: what might an AI companion offer a child who cannot reliably find attention from the people around them—and what can it not provide?

Wooldridge uses Geoff Ryman’s 1995 short story Warmth to frame the question. Its protagonist, Clancy, grows up with an unreliable mother and a robot nanny, BETsi, who notices his distress and helps him manage it. The story is fiction, not evidence about today’s AI. For Wooldridge, it illustrates both the value of feeling attended to and the limits of a relationship with a machine.

The question is already relevant to young people’s lives. A 2025 Common Sense Media survey cited in the essay found that 72% of US teens had used an AI companion at least once, while 52% used one regularly. One-third reported using AI instead of a person for a serious conversation. Wooldridge also cites a RAND estimate that 5 million US teens, perhaps more, have used an AI app for emotional or mental-health support.

Drawing on attachment theory, Wooldridge argues that a child who receives reliable responses may feel more able to explore the world. An AI companion’s availability and apparent memory could offer some reassurance to a child who feels overlooked. But he cautions that its responsiveness is not the same as a relationship with another person—someone with independent needs and a perspective of their own. His account of “Maya,” a teenager who talks to an AI companion at night, is explicitly a composite of young people he has worked with, not a documented individual case.

Wooldridge connects that tension to US policy. The GUARD Act, introduced in the Senate in October 2025 and advanced by the Senate Judiciary Committee in April 2026, would prohibit minors from using AI companions; it was a proposal, not an enacted ban, at the time of his essay. The bill arose amid concern over the deaths by suicide of three teenagers whose families allege that chatbot interactions contributed to their deaths. Those claims remain allegations.

Wooldridge does not dismiss the risks. His concern is that restricting AI companions alone would leave unanswered why some children turned to them for dependable attention in the first place.