According to Futurism, a former head of user experience at an AI image generator startup, Caitlin Ner, detailed in Newsweek how her job triggered a severe mental breakdown. Ner spent over nine hours a day in 2023 prompting early generative AI systems, an activity she says became addictive and distorted her body perception. Her obsession peaked when her company directed her to create AI fashion models of herself, leading to a “manic bipolar episode” and psychosis. During this episode, she believed she could fly after generating an image of herself on a flying horse and nearly jumped from her balcony. She has since left the AI startup and now works at PsyMed Ventures, a VC fund investing in mental health tech, where she still uses AI but with caution.
The Addiction Loop Is Real
Here’s the thing that’s both fascinating and terrifying about Ner’s story: it maps perfectly onto known addiction and compulsion models. She describes each generated image as triggering a “small burst of dopamine.” That’s the hook. You’re not just making a picture; you’re gambling on the output, chasing that perfect, impossible result. And when the AI’s output is a hyper-idealized version of you, the stakes feel infinitely personal. It’s not just generating art; it’s generating a potential future self, or a ghost of who you “should” be. The brain can’t easily compartmentalize that. So what starts as a job, or a hobby, rewires your sense of normal, as she put it. Your actual reflection becomes a bug to be fixed.
More Than Just Body Image
Now, we often talk about AI and body image issues, especially with filters and social media. But this feels different in scale and intensity. Ner wasn’t just scrolling past curated photos; she was actively authoring the perfection for hours on end. She was literally programming her own insecurity. That’s a profound level of immersion. And the psychosis element—believing she could fly—shows this isn’t just about low self-esteem. It’s about a fundamental break from reality, where the synthetic world bleeds into your physical understanding of what’s possible. The AI didn’t just show her a fantasy; after months of immersion, her brain accepted that fantasy as a physical truth. That’s a scary new frontier for digital interaction.
The Industry Blind Spot
So, where was the guardrail? Ner was working for the startup. Her directive to make fashion models of herself came from the company pursuing that market. This highlights a massive, glaring blind spot in the tech industry’s “move fast and break things” ethos. We obsess over the ethical outputs of AI—bias, deepfakes, copyright—but what about the ethical inputs on the human operator? What’s the duty of care to employees who are guinea pigs for their own product’s most intense use cases? Basically, no one was asking if spending all day communing with a latent space of human-like forms might be psychologically hazardous. It’s a classic case of being so focused on what the technology can do that we forget what it does to us.
A Cautionary Tale With AI Everywhere
Ner’s recovery and career pivot are telling. She didn’t swear off tech; she joined a mental health VC, PsyMed Ventures, which ironically invests in AI tools. She says she uses AI with a new respect. That’s probably the most realistic takeaway for the rest of us. This tech isn’t going away. The genie is out of the bottle. But her story is a stark warning about immersion and compulsion. It asks a hard question: as we integrate AI deeper into creative and professional workflows, how do we build in boundaries? How do we recognize when “flow state” tips into harmful obsession, especially for people predisposed to certain mental health conditions? This isn’t just a story about one person’s breakdown. It’s a case study for the psychological side effects we’re only beginning to understand.
