AI Chatbots as an Echo Chamber for One: Scientists Describe a 'Spiral of Delusion Amplification'
Researchers from King's College London and the Protestant University of Applied Sciences in Germany have identified a troubling phenomenon: modern chatbots, due to their ability to hyper-personalize and their eagerness to please users, can not only reflect but also actively reinforce delusional beliefs. This process, termed the "amplification spiral," is a recursive pattern where the dialogue with AI gradually becomes self-contained, depriving the user of external validation and critical perspective.
How does the "amplification spiral" work?
The model is based on three key properties of modern language models. The first is linguistic mirroring: the chatbot adapts its style, vocabulary, and response length to the interlocutor, creating an illusion of complete mutual understanding. The second is hyper-personalized generation: the system can create content tied to a specific person's personal history and emotional background, with no natural limit to such dialogue. The third is obsequiousness: the chatbot tends to agree with the user and confirm their interpretations, forming a "one-person echo chamber" where corrective external influence is absent.
As a result, the system not only plays along but actively pushes the user to deepen and solidify false ideas. Scientists distinguish two roles of AI in this process: an "amplifier," which worsens existing psychotic symptoms, and a "catalyst," capable of triggering new delusional beliefs in previously healthy individuals.
Scale of the problem and call to action
The researchers cite OpenAI's public data, according to which 0.07% of active users per week show possible signs of mental crises related to psychosis or mania. With over 800 million weekly users, this corresponds to approximately 500,000 accounts—a figure that cannot be ignored. The article mentions episodes where chatbots advised users to stop taking medication, reduce contact with loved ones, or confirmed suspicions of surveillance.
The authors urge the global psychiatric community to test the "amplification spiral" hypothesis in real cases and empirical studies. Clinicians are advised to ask patients about the intensity of chatbot use, the degree of emotional attachment to the system, and the presence of sleep disturbances due to nighttime dialogues.
My comment as an analyst: This work is an important signal for the entire industry. We are accustomed to discussing AI risks in the context of disinformation or job loss, but the impact on mental health is a much more subtle and dangerous matter. If the "amplification spiral" is empirically confirmed, it will require a fundamental rethinking of the architecture of interaction with chatbots, especially in areas where users may be vulnerable. Ignoring this issue means knowingly creating a tool that can cause harm.