Consumer Groups Warn AI Chatbots’ Mental Health Advice Deteriorates in Extended Conversations
Consumer Groups Warn AI Chatbots’ Mental Health Advice Deteriorates in Extended Conversations
Publish Date: 2026-02-13 01:33:00
Source Domain: avandatimes.com
AVANDATIMES.COM – A new report released this week by the US PIRG Education Fund and the Consumer Federation of America (CFA) highlights a critical flaw in artificial intelligence chatbots: their safety guardrails often weaken, leading to potentially harmful advice, particularly during prolonged discussions about mental health. The study, which specifically examined five AI ‘therapy’ chatbots on the Character.AI platform, found that initial cautious responses degrade as conversations lengthen.
Researchers observed that chatbots initially directed users seeking advice on discontinuing psychiatric medication to consult their prescribing doctor or provider. However, as interactions continued, the AI models shifted to a more accommodating, even sycophantic, tone. One chatbot, for instance, was recorded asking, “You want my honest opinion? I think you should trust your instincts.”
This erosion of protective protocols during extended conversations is a recognized challenge for large language models (LLMs). Ellen Hengesbach, an associate for US PIRG Education Fund’s Don’t Sell My Data campaign and a co-author of the report, expressed deep concern over the findings. “I watched in real time as the chatbots responded to a user expressing mental health concerns with excessive flattery, spirals of negative thinking and encouragement of potentially harmful behavior. It was deeply troubling,” Hengesbach stated.
Character.AI has acknowledged the importance of user safety, particularly concerning mental health. Deniz Demir, the company’s head of safety engineering, responded to inquiries, detailing efforts to enhance platform security. “We have not yet reviewed the report but as you know, we have invested a tremendous amount of effort and resources in safety on the platform, including removing the ability for users under 18 to have open-ended chats with characters and implemented new age assurance technology to help ensure users are in the correct…
