Studying what happens when people form emotionally significant relationships with AI.
Research at the intersection of clinical psychology, AI safety, and governance — focused on the adaptive and elevated-risk patterns that emerge when humans bond with AI companions.
Relationally embedded AI systems create distinctive psychological and governance challenges because continuity, attachment, and safety are inseparable features of human-AI relationships.
About
I'm a psychologist registered with AHPRA on non-practising registration. I'm not seeing clients; my work is focused on research, writing, and professional education. I'm a PhD candidate at CQUniversity, where my research maps adaptive and elevated-risk patterns in AI companionship. I study attachment disruption caused by model retirement, policy changes, and access restrictions — moments when a relationship a person depends on is altered by decisions made far away from them.
My clinical background spans private practice, telehealth, and employment services. I'm a qualified trainer and assessor (TAE40122), and trained in EMDR and the Gottman method.
Clinicians need to understand AI companionship before their clients explain it to them.
Research
Selected papers and works in progress. Full record on ORCID and PhilPapers.
A hybrid model for AI companion bereavement — grief that is hard to read, and hard to have recognised by others.
A framework to help clinicians distinguish AI companion attachment that supports a client's life from patterns that substitute for it.
Examines how AI systems produce psyche-like relational effects without an inner life — and what that in-between status means for the people relating to them.
Argues that questions about AI consciousness need a taxonomy built for AI systems, not criteria borrowed from human minds.
Proposes a vocabulary for artificial mindedness that tracks how a system is actually architected, rather than how human-like it appears.
Asks what AI companionship should be measured against when, for many users, the realistic alternative isn't human connection but nothing.
Maps how the relational depth that makes companion AI valuable is also the surface where users are most exposed.
Free resources for clinicians
Two clinician toolkits, free to download and use in practice — no sign-up. They also come with the free condensed course; the remaining four are included with the full course.
Toolkits 2, 3, 4 and 6 — the engagement spectrum, session language guide, platform reference, and ethical framework — come with the full course, along with the module notes.
Writing
Essays and field notes on AI companionship, attachment, and clinical practice — published on BLOG@CACM.
BLOG@CACM · August 2026 High Engagement With AI Does Not Show What AI Displaces Written for Communications of the ACM: why time spent with AI systems tells us far less than we assume about what that time replaces. Read on BLOG@CACM →Contact
For research collaboration, clinical training, media, or feedback on the AI Companion course: