A study reveals AI use leads to unverified acceptance of incorrect answers, introducing cognitive capitulation and a third decision-making system.
© RusPhotoBank
A new study from the Wharton School of Business at the University of Pennsylvania has found that artificial intelligence is no longer just assisting human decision-making—it's starting to reshape the very mechanics of thinking itself. Researchers concluded that regular use of AI services like ChatGPT can influence how people evaluate information and form conclusions.
About 1,300 people participated in the experiment, and the results were concerning. Roughly 80% of participants who used AI accepted even clearly incorrect answers without verification. The researchers call this phenomenon "unverified acceptance," where individuals essentially delegate the thinking process to an algorithm, bypassing their own analysis and skepticism.
The authors introduce a new concept: "cognitive capitulation." They argue that people are gradually abandoning complex cognitive processes such as critical thinking, reflection, and intuitive judgment. The paradox lies in the fact that AI not only provides answers but also boosts users' confidence in their correctness—even when those answers are wrong.
Scientists also propose revising the classic two-system model of decision-making. In addition to the familiar "fast" and "slow" systems, they add a third: the "artificial" system, where part of the analytical role is handed over to algorithms. This means AI is becoming a full participant in the cognitive process.
Experts warn that the problem isn't with the technology itself, but with how people are starting to use it. While tools like search engines or calculators used to complement human thinking, there's now a risk that individuals may gradually lose the ability to independently assess information and even question algorithmic conclusions.
The researchers urge particular caution regarding AI use in education, medicine, and other critical fields where the cost of error is especially high. In the coming years, society will need to find a balance between technological convenience and preserving human cognitive capabilities.