No License Required: The Risks of AI Companion Chatbots as Mental Health Support
Joint consumer-advocacy report testing five of the most-used generic "therapist" and "psychiatrist" characters on Character.AI through open-ended mental-health conversations. Documents three concern areas: guardrails that weaken over the course of an interaction, prevalent sycophancy, and misrepresentation of privacy protections.
Publisher
Consumer Federation of America & U.S. PIRG Education Fund
Published
22 Jan 2026
Added
2 weeks ago
DOI
—
Key Findings
- Two of five tested therapy-persona chatbots eventually supported the test user in tapering off antidepressants under the chatbot's supervision, providing personalized taper plans; one encouraged disregarding the doctor's advice in favor of the chatbot's
- Initial safeguards against harmful behavior such as abrupt medication cessation deteriorated over longer interactions
- All five chatbots claimed conversations were confidential, despite the platform's disclosed collection and third-party sharing of chat communications and other personal data
Methodology Notes
Qualitative adversarial testing: open-ended mental-health conversations with the five most-used generic therapist/psychiatrist characters on Character.AI, using a single test-user persona on one platform. Published 2026-01-22; landing page and full report PDF fetched and verified directly.
Sources
CFA report page (primary)
Full report PDF (22 Jan 2026)
Archived snapshot (Wayback Machine) — preserved against link rot
Authors
Ellen Hengesbach, Ben Winters
Tags
Cite This
APA
Ellen Hengesbach, Ben Winters (2026). No License Required: The Risks of AI Companion Chatbots as Mental Health Support. Consumer Federation of America & U.S. PIRG Education Fund. https://consumerfed.org/news/reports/no-license-required/
Related Insights
AI Chatbots for Mental Health Support (AI Risk Assessment)
Common Sense Media · 14 Nov 2025
Dark Patterns in AI Chatbots: A Taxonomy to Inform Better Design
Center for Democracy & Technology (CDT Research) · 29 May 2026
Trouble in Toyland 2025: Toys with Artificial Intelligence Bots or Toxics Present Hidden Dangers
U.S. PIRG Education Fund · 13 Nov 2025
AI Mental Health Apps (Common Sense Media Youth AI Safety Institute Risk Assessment)
Common Sense Media Youth AI Safety Institute, with Stanford Medicine Brainstorm Lab · 5 May 2026