CoChat

Providing Self-Led Mental Health Support Through an Artificial Intelligence–Powered Chat Bot (Leora) to Meet the Demand of Mental Health Care

Open access
van der Schyff, Emma L, Ridout, Brad, Amon, Krestina L, Forsyth, Rowena, Campbell, Andrew J · 2023 · Journal of Medical Internet Research

Published version of a preprint

The DOI matches its Crossref record; 41 of the 42 references it cites were found in a registry or an index; 17 of the 31 sources its claims cite were checked for support.

Abstract

Digital mental health services are becoming increasingly valuable for addressing the global public health burden of mental ill-health. There is significant demand for scalable and effective web-based mental health services. Artificial intelligence (AI) has the potential to improve mental health through the deployment of chatbots. These chatbots can provide round-the-clock support and triage individuals who are reluctant to access traditional health care due to stigma. The aim of this viewpoint paper is to consider the feasibility of AI-powered platforms to support mental well-being. The Leora model is considered a model with the potential to provide mental health support. Leora is a conversational agent that uses AI to engage in conversations with users about their mental health and provide support for minimal-to-mild symptoms of anxiety and depression. The tool is designed to be accessible, personalized, and discreet, offering strategies for promoting well-being and acting as a web-based self-care coach. Across all AI-powered mental health services, there are several challenges in the ethical development and deployment of AI in mental health treatment, including trust and transparency, bias and health inequity, and the potential for negative consequences. To ensure the effective and ethical use of AI in mental health care, researchers must carefully consider these challenges and engage with key stakeholders to provide high-quality mental health support. Validation of the Leora platform through rigorous user testing will be the next step in ensuring the model is effective. ©Emma L van der Schyff, Brad Ridout, Krestina L Amon, Rowena Forsyth, Andrew J Campbell. Originally published in the Journal of Medical Internet Research.

Overview

CoChat Summary

TL;DR

This viewpoint paper proposes Leora, an AI-powered conversational agent for self-led mental health support, and argues that it can help meet demand for scalable, discreet, and accessible support for minimal-to-mild anxiety and depression. The paper is primarily conceptual and ethical, emphasizing transparency, user safety, and the need for rigorous validation before broad deployment.

Key findings

  • Leora is described as a web- and app-based conversational agent that can assess, monitor, and manage minimal-to-mild symptoms of anxiety and depression, and provide evidence-based coping strategies and well-being support.
  • The proposed triage model uses standardized measures, specifically the GAD-7 and PHQ-9, to route users: those with clinically significant scores are referred to mental health services, while those with low/no symptoms receive psychoeducation and mindfulness strategies.
  • The authors state that Leora is not designed to assist with mental health crises and will direct users with abuse, self-harm, severe symptoms, or other emergencies to emergency mental health services.
  • The paper argues that AI chatbots may help address unmet demand because web-based mental health services are needed at scale, and cites prior chatbot examples: Woebot reportedly reduced depression and anxiety symptoms in young adults, and Wysa reported reduced self-reported depression symptoms and 68% of a cohort finding it helpful and encouraging.

Methods

  • This is a viewpoint paper; it presents a proposed model rather than original empirical testing.
  • The paper is based on conceptual design and literature-informed discussion of AI mental health chatbots, digital mental health, and ethical implementation.
  • Main analysis is a narrative argument describing Leora’s workflow, safety architecture, triage pathways, and ethical considerations, with references to existing tools and prior studies.

Limitations

  • The paper explicitly says validation of the Leora platform through rigorous user testing will be the next step, so the model is not yet empirically validated.
  • The text notes that, more broadly, evidence for conversational agents is currently weak due to the limited number of clinical trials and high estimated risk of bias in the evidence base.
  • Evident from the paper’s framing: it is a proposed model / viewpoint, so claims about effectiveness, acceptability, and clinical outcomes are not supported by results from Leora itself.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.2196/46448

Cites

41 of 41 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1 Road map for self-assessment of mental ill-health on the Leora platform.
Figure 1 Road map for self-assessment of mental ill-health on the Leora platform.CC BY · PubMed Central
Figure 2 Assessing the effectiveness of the mental health support strategies offered by Leora.
Figure 2 Assessing the effectiveness of the mental health support strategies offered by Leora.CC BY · PubMed Central
Figure 3 Leora platform interface: chatting to the conversational agent, self-assessment, and booking therapy.
Figure 3 Leora platform interface: chatting to the conversational agent, self-assessment, and booking therapy.CC BY · PubMed Central

Verification

Claims the paper makes

14 Valid · 7 Not valid · 9 Not checked · 13 Own results and interpretations

9 more claims · 2 with their cited sources checked

Create a free account to see all 43 claims and every source verdict.

See all claims — free account
  • It may state two separate claims as one. 19
  • Probably not a substantial claim. 23

Create a free account to see these claims.

Citations

Cites

Show all 41 references

Research works printed but not linked: 2

  1. Naeinian, MR, 2011, Clin Psychol Pers, 9, 41
  2. 10.3316/informit.545561433272993 · doi:10.3316/informit.545561433272993

Cited by

Metadata

Authors
van der Schyff, E. L., Ridout, B., Amon, K. L., Forsyth, R., & Campbell, A. J.
Affiliations
The University of Sydney (AU)
Venue
Journal of Medical Internet Research
Published
2023-06-19
Volume
25
Pages
e46448
ISSN
1438-8871 (ISSN-L)
OpenAlex
W4377084726
Cited by
109
References
42
Citations by year
2026: 34 · 2025: 53 · 2024: 19 · 2023: 3
FWCI
28.05
Type
journal article
Language
en
Open access
Licence
cc-by
Locations
  • Journal of Medical Internet Research (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Topics
  • Social Sciences › Psychology › Applied Psychology › Digital Mental Health Interventions
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
  • Health Sciences › Medicine › Psychiatry and Mental health › Mental Health Treatment and Access
Keywords
chatbots, digital mental health, anxiety, depression, mental health stigma, ethical use of AI, algorithmic bias, health disparities, trust and transparency, mental health demand
MeSH terms
Psychological Well-Being, Artificial Intelligence, Humans, Mental Health, Mental Health Services, Acceptance and Commitment Therapy

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-11, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-11 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.