Thinking, Fast, Slow, and Artificial

Here is what NotebookLM said about this paper:

This text introduces Tri-System Theory, a new cognitive framework that adds System 3 (artificial cognition) to the traditional model of fast and slow human thinking. Through three large-scale experiments, the researchers demonstrate cognitive surrender, a phenomenon where individuals abandon their own reasoning to uncritically adopt AI-generated answers. The findings reveal that while accurate AI can boost performance and buffer the negative effects of time pressure, faulty AI significantly decreases accuracy because users fail to verify its outputs. This tendency to surrender is most common among people with higher trust in AI and a lower personal drive for effortful thought. Although performance incentives and feedback can encourage more critical oversight, the study concludes that AI fundamentally reshapes human autonomy by displacing internal deliberation. This research highlights the shift from internal biological reasoning to a triadic cognitive ecology where humans and machines co-produce judgment.

Suggested citation:

Shaw, Steven D and Nave, Gideon, Thinking—Fast, Slow, and Artificial: How AI is Reshaping Human Reasoning and the Rise of Cognitive Surrender (January 11, 2026). The Wharton School Research Paper , Available at SSRN: Link to the SSRN page for the article .

On April 15, 2026, GCC's AI Forum will listen to this podcast debate creating by NotebookLM and discuss it: