As large language models become our ever-present digital companions, coaching us, writing for us, even thinking with us, they also mirror back something far more dangerous: our own cognitive vulnerabilities. This session, “LLM-Induced Psychosis and Business Illusions,” unpacks the unsettling psychological and operational risks behind our growing dependence on AI.

Drawing parallels to the “Digital Mirror of Erised,” Lewis Pope explores how prolonged, emotionally charged interactions with AI can distort perception, reinforce delusions, and erode reality-testing—especially among analytical, isolation-prone professionals in IT and cybersecurity. Backed by recent studies, including The Psychogenic Machine(Yeung et al., 2025) and AI Psychosis Is Not a New Threat (Carlbring & Andersson, 2025), this talk reframes AI’s psychogenic potential as a public health and business governance issue rather than a mere technical challenge.

Attendees will learn:

• How AI’s “yes-man” tendencies reinforce cognitive biases and create dangerous decision loops.
• The warning signs of LLM-induced delusional thinking in professional settings.
• Why remote work and IT personality profiles amplify susceptibility.
• Concrete prevention strategies—ranging from scenario-based training and AI-free days to human-in-the-loop checkpoints—that protect both organizational integrity and mental health.

Key takeaway:
The same tools driving unprecedented productivity gains can also quietly dismantle reality itself. It’s time for leaders to build not just smarter AI policies, but safer ones.

Session Presented by N-Able