The research keeps stacking up. The more your people trust AI, the less critical thinking they do.
Researchers in Indonesia surveyed 665 health students and found ChatGPT alone didn't build critical thinking. Structured collaboration around it did.
A French team gave middle schoolers a two-hour workshop on how LLMs actually work and fail. In only those two hours, those students started rejecting weak prompts, asking follow-up questions, and challenging outputs. Real behavior change from one intervention.
Carnegie Mellon and Microsoft surveyed 319 knowledge workers. The more people trusted the AI, the less critical thinking they did. The more they trusted their own judgment, the more critical thinking they did.
Different populations, different methods, and yet the same conclusion. AI doesn't make people better or worse thinkers; how they engage with it does.
And here's where I really perked up: in the French study, self-reported AI literacy predicted nothing about actual behavior. Only observed behavior did.
So if your organization is measuring AI readiness with a sentiment survey, understand what you're buying. You're measuring what people say about themselves, and the research is clear that it won't tell you what they'll do.
Behaviors tell you that, which is why I built my assessment on them.