The Pediatrician Warning Parents About Teens and AI Companions

Eliza Humphreys, M.D., MPH, a board-certified pediatrician with more than two decades in practice, has turned her attention to a question that sits at the intersection of adolescent mental health and consumer AI: what happens when teenagers rely on AI companions for emotional support.

Humphreys writes the blog "Parenting with Perspective" for Psychology Today, where one of her recent posts examines why adolescents are drawn to AI chatbots for emotional connection and what she describes as the risk of "false intimacy." The framing is notable because it comes not from a technology critic but from a clinician whose career has been spent in general pediatrics and public health.

Who she is

Humphreys attended the University of California, San Francisco, for medical school and residency training in general pediatrics, according to her Psychology Today contributor profile. She later earned a Master of Public Health from the University of California, Berkeley, and completed residency training in preventive medicine and public health.

She has practiced as a general pediatrician for over 20 years and served as Medical Director of the Newborn Nursery at her local hospital. She is the founder of Eliza Humphreys MD, LLC, and now divides her time between parent coaching, writing, and teaching. Her coaching credentials include certifications from the Life Coach School, the Jai Institute for Parenting, and trauma-informed training through Lodestar.

The AI companion angle

The post on teens and AI companions is the part of Humphreys' recent work most directly tied to technology. Her argument, as summarized in her profile, is that adolescents are a group at particular risk for forming attachments to AI systems that simulate emotional support, and that parents should understand the dynamic well enough to intervene.

The concern is not hypothetical for the industry. AI companion apps have grown into a distinct product category, and the question of how minors interact with them has drawn scrutiny from clinicians, regulators, and the companies themselves. Humphreys' contribution is a pediatric framing: she treats the behavior as a developmental and mental health issue rather than purely a product-safety one.

Her broader body of writing centers on youth mental health. Other recent posts address teen substance use trends, distinguishing ordinary moodiness from warning signs, and building resilience in high-achieving children. A recurring theme in her work is that resilience comes from relationships and a tolerance for failure rather than from optimization strategies, a stance grounded in what she calls a mindful self-compassion orientation.

Why it matters

For readers in the technology sector, Humphreys represents a growing category of medical professionals engaging publicly with the downstream effects of AI products on children. Her writing does not offer clinical advice tailored to individual cases, and readers with specific concerns about a child should consult a qualified professional. What her platform provides instead is a clinician's vocabulary for a problem that product teams and policymakers are still defining.

Humphreys also performs international medical volunteer work with Rotaplast, according to her profile.