Back to search

Article

AI Safety Training Can be Clinically Harmful

2026-06-12

Abstract excerpt

<title>Abstract</title> <p>Large language models are being deployed as mental health support agents at scale, yet only 16\% of LLM-based chatbot interventions have undergone rigorous clinical efficacy testing \citep{hua2025charting}. We argue that the dominant alignment paradigm, RLHF-based safety training, can itself become a source of clinical harm when models are placed in protocol-driven psychotherapeutic set...

Topics

Open a Topic to create a Post that cites this publication.

Identifiers and source

Literature Corpus work
ed499ba0-009d-529e-9c49-ed9c0b65fd8f
DOI
10.21203/rs.3.rs-9568253/v1
Open publication

Related research

Semantic proximity does not establish scientific evidence.

Click a neighbor to travelStep 1 · 12 closest
Interactive article relationship graphSelect a related publication card to move it into the centre and load its closest explainable connections. Solid lines are source-backed structured connections. Dashed lines are semantic discovery signals and are not scientific evidence.
AI Safety Training Can be Clinically HarmfulDOI 10.21203/rs.3.rs-9568253/v1
ΔAPT: Can we build an AI Therapist? Interdisciplinary critical review aimed at maximizing clinical outcomes using Large Language Models for AI Psychotherapy.Semantic proximity
Safety Mechanisms and Risk Mitigation in Generative AI Mental Health Chatbots: A Systematic Scoping ReviewSemantic proximity
The Role of Large Language Models in Clinical Psychology: Current Applications, Efficacy, Challenges, and Future DirectionsSemantic proximity
A Scoping Review of Generative AI in Mental Health SupportSemantic proximity
A Scoping Review of Generative AI in Mental Health SupportSemantic proximity
AI in Psychotherapy: Opportunities and RisksSemantic proximity
Safety Evaluation of a Clinical-Grade Generative AI Agent for Anxiety and Depression SymptomsSemantic proximity
AI-Driven Mental Healthcare 5.0: A Survey of Opportunities and Challenges in Leveraging Large Language Models and Generative AISemantic proximity
Large Language Models Acting as Psychotherapists: A Scoping ReviewSemantic proximity
Evaluating and Enhancing an AI Mental Health Guide in High-Risk Mental Health ContextsSemantic proximity
Risk Governance for Generative AI Mental Health Support: A Multi-Turn Safety ArchitectureSemantic proximity
Safety Evaluation of a Generative AI Agent for Anxiety and Depression SymptomsSemantic proximity
Select a neighboring publication to make it the new centre.