AI psychosis

Definition: over-trust in AI output (the state where a user becomes convinced AI is more capable than it is, typically after repeated “one-shot automation awe” experiences).

Origin

Term from 2026-10-04-youtube-ai-skeptic-sabbath (Arabic-transcript YouTube). The speaker self-identifies as a skeptic who entered this state: “I trust AI more than myself. I’m not suffering from AI psychosis. I’m not a promoter. But I trust AI more than I trust myself.”

Mechanism

1. One-shot automation awe: throwing a feature list at AI and walking away, you return amazed it worked; the pattern then repeats
2. Compounding false confidence: repeated successes create illusion that AGI has arrived
3. Skill atrophy coupling: over-trust leads to under-practice, which deepens dependence

Contrast

Not the same as enthusiasm or adoption. Psychosis implies a break from evidence. The speaker explicitly acknowledges their position “isn’t purely evidence-based” and cites studies on this phenomenon.

Related

• skill-atrophy-ai: the coupled degradation from disuse
• one-shot-automation-awe: the triggering experience
• slop-code-bench: evidence that the trust is misplaced for iterative work