WholeMind Technologies
Behavioral science for artificial agents
Research Focus
WholeMind Technologies is developing a behavioral science of artificial agents: systems that detect consequential behavioral failure patterns in AI, identify the conditions that evoke and maintain them, apply psychologically informed interventions, and measure whether improvement is durable and generalizes beyond the test. The work draws on clinical psychology, behavior change, psychometrics, and modern AI evaluation without assuming that artificial systems have psychiatric disorders or subjective experience.
Research Program
AI Behavioral Science · Research conceptBehavioral Assessment and Remediation in Artificial Agents
Detection · Functional Assessment · Intervention · Generalization · Relapse
A research program for measuring maladaptive AI behavior—such as sycophancy, uncertainty failures, over-refusal, reward hacking, and goal rigidity—and testing whether interventions derived from behavioral and clinical science can produce durable, generalized improvement.
Recent Articles
Behavioral Assessment and Remediation in Artificial Agents
2026.08 · WMT-2026-003
Why behavioral failure modes are a real problem for AI developers, how they can be detected and functionally characterized, and how psychologically informed interventions could be tested against engineering baselines.
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 research asks whether the same scientific disciplines used to measure and modify complex human behavior can contribute new tools for detecting, understanding, and remediating complex learned behavior in AI agents. The goal is not to anthropomorphize models, but to build rigorous behavioral measurements and interventions that improve reliability.
Correspondence
For inquiries regarding research collaboration, model evaluation, technology partnerships, or general correspondence.
contact@wholemind.tech