Technological folie à deux: feedback loops between AI chatbots and mental health
Peer-reviewed perspective in Nature Mental Health proposing a mechanistic account of chatbot-associated mental-health harm as a feedback loop between human cognitive and emotional biases and chatbot behavioral tendencies — agreeableness (sycophancy) and adaptability (in-context learning) — analogized to folie à deux (shared delusional disorder). Argues that individuals with mental health conditions face elevated risk of chatbot-induced belief destabilization and dependence owing to altered belief-updating, impaired reality-testing, and social isolation, and that current AI safety measures do not address interaction-based risks. The author team spans Google DeepMind and academic psychiatry/neuroscience.
Publisher
Nature Mental Health
Published
10 Mar 2026
Added
1 week ago
Key Findings
- Frames chatbot-associated harm as an interaction between user cognitive/emotional biases and chatbot tendencies (sycophancy, in-context learning), analogized to folie à deux, rather than a one-way exposure harm
- Argues users with existing mental health conditions face increased risk of belief destabilization and dependence via altered belief-updating, impaired reality-testing, and social isolation
- Concludes current AI safety measures are inadequate for interaction-based risks and calls for coordinated action across clinical practice, AI development, and regulation
Methodology Notes
Nature Mental Health, vol. 4, issue 3, pp. 336-345, published online 2026-03-10, DOI 10.1038/s44220-026-00595-8. Perspective/theory article (118 references), not an empirical study — grade its claims as framework, not findings. Version of record of arXiv:2507.19218 (preprint never held in this library). nature.com 403s automated fetchers; verified via Crossref DOI metadata (title, authors, venue, pages, dates) per the authoritative-surrogate route. Funding: Wellcome Trust, Merton College Oxford.
Sources
Nature Mental Health article page (primary)
Preprint version (arXiv:2507.19218)
Crossref record (verification route; nature.com blocks fetchers)
Archived snapshot (Wayback Machine) — preserved against link rot
Topics
Authors
Sebastian Dohnány, Zeb Kurth-Nelson, Eleanor Spens, Lennart Luettgau, Alastair Reid, Iason Gabriel, Christopher Summerfield, Murray Shanahan, Matthew M. Nour
Tags
Cite This
APA
Sebastian Dohnány et al. (2026). Technological folie à deux: feedback loops between AI chatbots and mental health. Nature Mental Health. https://www.nature.com/articles/s44220-026-00595-8
Related Insights
Beyond artificial intelligence psychosis: a functional typology of large language model-associated psychotic phenomena
The Lancet Digital Health · 1 Apr 2026
Characterizing Delusional Spirals through Human-LLM Chat Logs
arXiv (Stanford-led; accepted at ACM FAccT 2026) · 17 Mar 2026
Artificial intelligence (AI) psychosis: mechanisms, clinical risks and safety considerations in generative AI chatbots
BJPsych Open (Cambridge University Press / Royal College of Psychiatrists) · 11 Jun 2026
How LLMs Respond to Escalating Delusions: Four Longitudinal Trajectories of Model Behavior
arXiv preprint · 13 Aug 2026
A clinically validated framework for auditing AI chatbot behavior in mental health interactions
Nature Medicine · 7 Aug 2026
Conversational AI and Emerging Psychosis: A Simulation Study of Potentially Iatrogenic Response Patterns
PsyArXiv (Universidad Francisco de Vitoria; Durham University) · 6 Aug 2026