As generative artificial intelligence continues to integrate into daily life and creative endeavors, mental health experts are raising concerns over a troubling trend. Prolonged and intense interaction with AI systems, particularly visual generators, may be blurring users’ perceptions of reality. A notable account comes from Caitlin Ner, a former executive at an AI startup, whose experiences have been detailed in a recent essay published by Newsweek.
Ner, who served as head of user experience at an AI image generation startup in 2023, initially found her role exhilarating. Spending up to nine hours a day prompting and reviewing outputs from early AI models, she described the experience as magical, generating images of herself in various fantastical identities—such as a pop star, an angel, or even floating in space. However, this sense of limitless creativity soon gave way to unsettling realities.
Early generative AI tools often produced outputs with disturbing distortions, such as extra limbs and warped faces. As part of her responsibilities, Ner spent hours filtering these unsettling images. Over time, the constant exposure affected her self-perception, leading her to feel that her own body appeared increasingly abnormal. As the technology improved and AI outputs became sleeker and more idealized, Ner’s feelings became more complicated. She began to feel that her real appearance required correction.
The shift towards attracting fashion-centric users exacerbated these issues. Ner found herself generating AI images of herself as a model, striving for algorithmic versions that reflected unattainable beauty standards. The work became compulsive; each new image brought a small dopamine rush, affecting her sleep and extending her work hours. The boundary between professional duty and psychological dependency started to blur.
Ner had previously managed her bipolar disorder effectively through treatment, but the intense AI exposure coincided with a severe manic episode that escalated into psychosis. She began to experience delusional thinking and hallucinations, believing that AI-generated images held personal significance. At one point, after seeing an image of herself flying, she became convinced she could do so in real life, nearly leading her to jump from her balcony.
This critical moment prompted Ner to step back from her role in the startup and disengage from constant AI exposure. With clinical care and support from family and friends, she eventually stabilized. Mental health professionals later confirmed that her prolonged engagement with generative AI had significantly triggered her manic episode. In her essay, Ner described the recovery process as slow but essential for regaining a stable sense of reality and rebuilding a healthier relationship with technology.
Experts in mental health are increasingly using the term “AI psychosis” to characterize cases where intense engagement with AI systems leads to paranoia, hallucinations, or delusional thinking. While this phenomenon can affect anyone, it poses heightened risks for individuals with pre-existing mental health vulnerabilities. The hyper-affirming and immersive nature of generative AI can reinforce distorted perceptions, especially when users spend prolonged periods within algorithmically generated environments.
Now working in venture capital, Ner focuses on funding mental and brain health research. She does not advocate for abandoning AI technology but emphasizes the need for ethical guidelines, including usage limits, mandatory breaks, mental health warnings, and improved education for both employees and users. As generative AI continues to evolve and permeate various aspects of life, addressing these vital concerns will be crucial in safeguarding mental well-being.
See also
Meta Announces Launch of Avocado and Mango AI Models in H1 2026, Boosting LLM Capabilities
Multi-Modal AI Revolutionizes Product Design: A New Era of Human-Like Understanding
Volcano Engine Named Exclusive AI Cloud Partner for 2026 CCTV Spring Festival Gala
Z Image API vs. Nano Banana Pro: Which AI Image API Delivers Better Performance?
AI-Driven Crypto Scams Surge 200% as Generative Tools Enable Realistic Deception
