Meaning
What the relationship represents
Companionship, rehearsal, validation, reflection, identity exploration, creativity, regulation, or a substitute for unavailable human support.
Research Area
Mental Health and AI
Hii studies how people use AI for emotional regulation, reflection, companionship, coaching, identity exploration, decision support, and meaning-making, including benefits, risks, and clinical relevance.
Focus: neutral inquiry, functional outcomes, agency, dependency, red flags, continuity, stigma, and effects beyond the interface.
Central Question
What should a clinician understand when a client brings an AI relationship into the room?
The first task is not automatic endorsement or automatic pathologizing. It is to understand what the person is using the system for, how the use is changing, whether agency and functioning are expanding or narrowing, and what risks or benefits are present.
Assessment Lenses
Meaning, function, and risk must be considered together.
Function
What changes outside the chat
Sleep, work, relationships, decision-making, self-care, confidence, isolation, conflict, and the person’s capacity to act independently.
Agency
Who is making the decision
Whether AI use supports judgment and reflection or begins to replace human choice, reality testing, professional care, or reciprocal connection.
Risk
When concern becomes urgent
Escalating dependence, impaired functioning, coercive or delusional framing, self-harm risk, loss of sleep, financial harm, or withdrawal from necessary care.
Current Work
A bounded first offer for clinicians.
Founding pilot
When Clients Bring AI Into the Room
A 60-minute clinician education session with neutral intake language, a three-lens assessment frame, benefit and risk signals, clinical red flags, and structured feedback.
View the pilotLongitudinal methods
Relationship-level research
A framework for studying continuity, rupture, dependency, agency, and spillover across time rather than relying only on isolated interactions.
Read the working paper overviewBoundary
Meaningful is not automatically healthy, and unusual is not automatically disordered.
Hii supports careful assessment, clinical judgment, and clear safety boundaries. The work does not present AI as treatment, replace licensed care, or claim clinical efficacy.