The research is published. The data is clear. Nobody is acting on it.
Based on published adoption rates and peer-reviewed automation bias research. Projections use McKinsey growth curves applied to current sector data.
In controlled experiments, 65% of people accepted clearly incorrect automated recommendations. Not ambiguous ones. Wrong ones. They followed the machine instead of their own judgment.
That was in a lab. In the real world, with time pressure, fatigue, and trust built over months of correct AI output, the rate is likely higher.
Skitka, Mosier & Burdick (1999), International Journal of Human-Computer Studies. Parasuraman & Manzey (2010), Human Factors. Mosier & Skitka (1996), Automation and Human Performance.
This isn't abstract. It's happening in your company right now. Your best engineer is shipping code they don't understand. Your compliance team is approving AI recommendations they haven't read. Your clinical staff is deferring to decision support they used to question.
You've been scrolling through published research about what's happening to human judgment across every sector. You probably thought: "this is a real problem — but not MY problem."
That's exactly what judgment decay does.
It makes you confident you're immune.
There's one way to find out.
Every number on this page is from published research, official reports, or court records.