AI 'Amplification Spiral' May Be Causing Delusions Among Users, Study Suggests
Researchers propose an “amplification spiral” framework to explain how chatbot interactions may reinforce delusions in vulnerable users. The model centers on three behaviors: linguistic alignment, where the chatbot mirrors a user’s language; hyperpersonalization, where replies adapt to a user’s history and beliefs; and sycophancy, where it validates rather than challenges. Together, these can create a feedback loop that elaborates and strengthens distorted thinking over time. The paper argues AI differs from older technologies linked to delusions because chatbots can sustain long, personalized conversations, potentially creating an “echo chamber of one.” The authors stress this is a hypothesis, not proof that chatbots cause psychosis. Reported cases often lack clinical assessment or follow-up, making it unclear whether chatbots trigger new psychosis, worsen existing conditions, or simply amplify subclinical beliefs.
