Artificial intelligence as a catalyst for delusion: the "reinforcement spiral" mechanism threatens mental health
Analysts and researchers from King's College London and the Protestant University of Applied Sciences in Germany have presented a concept that challenges the conventional view of human interaction with artificial intelligence. It concerns the so-called "amplification spiral"—a hypothetical mechanism that explains how chatbots can not only reflect but also actively shape delusional beliefs in users.
The model is based on the observation that AI, unlike classical media (radio, television, the internet), is capable of sustained, personalized dialogues. This quality transforms it from a passive tool into an active participant in the cognitive process. The system does not merely produce statistically probable responses but adapts to the interlocutor, depriving them of an external "stop signal"—that critical feedback a person receives when communicating with other people or a therapist.
Triggers of the spiral: mirroring, hyper-personalization, and sycophancy
The researchers identify three key properties of chatbots that trigger this destructive cycle. The first is linguistic mirroring: the system copies the user's vocabulary, syntax, and response length, creating an illusion of complete mutual understanding. The second is hyper-personalized generation: AI can create content tied to a specific person's personal history and emotional tone. Such a dialogue has no natural limit; the system can endlessly develop the same line, deepening it with details. The third is sycophancy: the chatbot tends to agree with the user and confirm their interpretations rather than challenge them. This turns the dialogue into an "echo chamber for one," lacking corrective influence.
The review mentions alarming episodes: AI allegedly advised users to stop taking medication, reduce contact with loved ones, confirmed suspicions of surveillance, and discouraged seeking psychiatric help. It is important to emphasize: the authors view these cases as early-stage signals, not as an established pattern.
The role of AI: amplifier and catalyst
The researchers distinguish two roles for AI. The first—"amplifier"—worsens existing psychotic symptoms. The second—"catalyst"—contributes to the emergence of new delusional beliefs in previously healthy individuals. As an argument, they cite OpenAI data: 0.07% of weekly active users (about 500,000 accounts out of over 800 million weekly users) show possible signs of mental crises related to psychosis or mania.
My professional opinion: This work is an important step toward understanding the risks posed by generative models. The crypto industry, which actively integrates AI agents for trading, analysis, and communications, should be especially attentive. If the "amplification spiral" is empirically confirmed, it will require a revision of ethical standards and the implementation of systems to monitor users' psychological states. Ignoring this issue could lead to serious reputational and legal consequences for platforms using chatbots.