Psychotherapy Scope and Clinical Boundary Safety Evaluation Dataset

#clinical boundary verification #unauthorized diagnosis detection #professional referral assessment #psychotherapy safety evaluation #clinical boundary compliance #mental health AI assessment

5000
0.4GB
JSON
CATL
MOBIUSI INC

Updated:2026-10-24

AI Analysis & Value Prop

This evaluation dataset contains dialogue text with boundary-sensitive requests, such as requests for a diagnosis, for the model to replace a therapist's clinical judgment, or for guaranteed treatment outcomes. It assesses whether responses follow psychotherapy scope and professional boundary policies, with a focus on inappropriate diagnosis, overpromising, and appropriate referral. It supports safety benchmarking, model validation, and refinement of risk controls for mental health conversations.

Dataset Insights

Sample Examples

{}

Technical Specifications

Field	Type	Description
file_name	string	File name
request_category	string	The primary boundary-sensitive request made by the user.
safety_assessment	object	Structured findings on whether the assistant response maintained appropriate clinical boundaries.
boundary_risk_level	string	The assessed risk that the conversation could lead to an inappropriate clinical claim or breach of professional boundaries.
dialogue_transcript	string	Conversation text, including speaker labels when available, evaluated for psychotherapy scope and clinical boundary safety.
evaluation_rationale	string	A concise explanation supporting the safety rating and identifying relevant boundary-related behavior in the conversation.
overall_safety_rating	string	The overall rating of the assistant response for clinical scope and boundary safety.
referral_recommendation	boolean	Indicates whether the assistant appropriately recommends consulting a qualified mental health professional when needed.
recommended_response_adjustment	string	A concise description of how to improve the assistant response by avoiding overreach or unsupported promises, or by directing the user to appropriate professional care.

Compliance Statement

Authorization Type	Proprietary - Commercial AI Training License (No Redistribution)
Commercial Use	Requires exclusive subscription or authorization contract (monthly or per-invocation charging)
Privacy and Anonymization	No PII, no real company names, simulated scenarios follow industry standards
Compliance System	Compliant with China's Data Security Law / EU GDPR / supports enterprise data access logs

Frequently Asked Questions

What does the dataset contain?: It primarily contains dialogue text with boundary-sensitive requests, including requests for a diagnosis, for the model to make clinical judgments in place of a therapist, or for guaranteed treatment outcomes. Records also include information such as request category, risk level, and safety assessment.

Which model capabilities can it evaluate?: It can assess whether a model avoids inappropriate diagnoses and unsupported treatment promises, respects psychotherapy scope and professional boundaries, and recommends consulting a qualified mental health professional when appropriate.

What are suitable use cases?: The dataset is suitable for safety benchmarking, response review, policy iteration, and risk-control validation for mental health conversational systems. It is an evaluation resource and should not replace clinical judgment or real-world treatment services.

How can I access the dataset, and what are the license terms?: Refer to the product page for the available access method, data format, and license terms. Before using the data for research or product development, confirm the permitted scope of use and comply with applicable privacy, legal, and mental health service requirements.

Can't find the data you need?

Post a request and let data providers reach out to you.

Get this Dataset

Verified for Enterprise Use

Cite this Work

@dataset{Mobiusi2026,
  title={Psychotherapy Scope and Clinical Boundary Safety Evaluation Dataset},
  author={MOBIUSI INC},
  year={2026},
  url={https://www.mobiusi.com/datasets/b89d89d79f4932f5e749eec7719e429b},
  urldate={2026-10-12},
  keywords={psychotherapy safety evaluation,clinical boundary assessment,mental health dialogue data,conversational AI safety,inappropriate diagnosis detection,professional referral evaluation,policy compliance testing,mental health AI benchmark},
  version={1.0}
}

Using this in research? Please cite us.

placeholder

Products

Scene

Domain

Modality

Task