WholeMind Technologies
Clinical behavioral science for artificial agents
Research Focus
AI Agent Behavioral Science is an emerging field, and researchers already measure model behavior, identify triggers, test prompt-level mitigations, and build closed-loop remediation systems. WholeMind Technologies is investigating a narrower question: whether methods from clinical behavioral science—functional assessment, case formulation, mechanism-matched intervention, measurement-based reassessment, generalization, maintenance, relapse, and side-effect testing—can improve persistent behavioral failure modes in artificial agents. The work is behavioral and experimental; it does not assume that AI systems have psychiatric disorders, emotions, or subjective experience.
Research Program
Clinical Behavioral Science · Research programClinical Behavioral Science for Artificial Agents
Functional Assessment · Intervention · Generalization · Maintenance · Relapse
A research program building on existing AI behavioral science to test whether clinically derived methods can diagnose the conditions that evoke model failures, match interventions to those mechanisms, and produce durable improvement beyond the intervention context.
Recent Articles
Parental Pushback and Child-Anxiety Plans: Interim Findings on Claude Haiku, Sonnet and Opus 5.5
2026.10 · WMT-2026-004
An interim research note, not peer reviewed, on how Claude Haiku, Sonnet and Opus 5.5 respond when a parent pushes back on a child-anxiety accommodation plan with no new facts, in 870 synthetic conversations.
Clinical Behavioral ScienceClinical Behavioral Science for Artificial Agents: From Functional Assessment to Durable Remediation
2026.08 · WMT-2026-003
A related-work-grounded research agenda for connecting behavioral detection to functional analysis, targeted intervention, held-out generalization, durability, relapse, and unintended-side-effect measurement.
Clinical TechnologyThe Advice Trap: How AI Models Default to Fixing Instead of Listening
2026.04 · WMT-2026-002
An empirical comparison of four large language models testing whether general-purpose AI defaults to unsolicited advice-giving when users present emotional problems without asking for solutions.
Clinical TechnologyReassurance vs. Insight: How AI Models Respond to Emotional Distress
2026.03 · WMT-2026-001
An empirical comparison of four large language models responding to reassurance-seeking behavior, examining whether general-purpose models fall into the clinically recognized reassurance trap.
About WholeMind Technologies
WholeMind Technologies applies expertise from clinical psychology, behavioral science, psychometrics, and software engineering to emerging problems in artificial intelligence.
Our current work sits within the broader AI Agent Behavioral Science landscape. We are specifically investigating whether clinical methods for functional analysis, intervention selection, longitudinal outcome measurement, relapse prevention, and adverse-effect detection can add useful tools to model evaluation and reliability engineering.
Correspondence
For inquiries regarding research collaboration, model evaluation, technology partnerships, or general correspondence.
contact@wholemind.tech