The Cost of Cognitive Surrender

In CAI Currents, The American Psychoanalyst, October 26, 2026

Should I or shouldn’t I? AI tells us it can reduce the frustration of hard problems, increase productivity, and make us look smarter and more articulate than we already think we are. But is it a siren’s call? When does use become a costly surrender?

Guy Levi, an innovator who focuses on developing universal Generative AI Literacy, framed the issue: “When its responses are fluid, convincing, and fast, the line between thinking with the model and thinking through the model can blur. We may begin to mistake fluency for insight, confidence for accuracy, coherence for truth. We may stop inspecting what it reflects. We may stop thinking altogether.”

We already have empirical evidence of “cognitive debt” resulting from AI use. A study at MIT found that “Over four months, LLM users consistently underperformed at neural, linguistic, and behavioral levels. These results raise concerns about the long-term educational implications of LLM reliance and underscore the need for deeper inquiry into AI’s role in learning.”

Gideon Nave and Steven D. Shaw, researchers at the Wharton School, studied the proclivity to offload thinking to AI in a paper that names the result: cognitive surrender, the uncritical acceptance of AI’s answers in place of one’s own reasoning.

As Shaw describes it, AI is “an expert across all sorts of different domains, and we have access anytime we want it. It speaks with confidence, carries few markers of doubt, and rarely says a subject is outside its area. It is not the advice of a fallible human mind. It is something new.”

The damage shows up in their experiments. When the AI gave confident wrong answers, participants did worse than people with no AI at all; their confidence went up anyway. Those who trust AI more are more likely to surrender. People who are more analytically inclined, who enjoy thinking problems through, are better protected.

The researchers focused on remedies within organizations: incentives for accuracy, training in using AI judiciously, and interfaces built to prompt users to pause and verify.

What is the takeaway for the rest of the population that is not governed by systems that have a requirement for accuracy? What is needed to address the power of AI to overtake critical thinking? It seems we are increasingly being called upon to protect our minds.