Even 10 minutes of using AI tools can significantly reduce our persistence in thinking

Researchers conducted a series of experiments involving 1,222 participants to check how the use of artificial intelligence affects persistence and efficiency in problem solving. In one of the key tests, a group of 354 people had to solve 15 fraction problems. Some participants could use ChatGPT, a popular AI tool that provided them with suggestions and ready-made answers. The second group had to rely solely on their own skills.

Initially, people using AI performed better, solved tasks faster and more accurately. However, after 12 questions, the researchers took away their access to the tool. The effect was immediate: the number of correct answers dropped dramatically, and participants began to skip questions more often. People who did not have AI support from the beginning solved the tasks until the end and with greater success. A similar pattern repeated in subsequent experiments, both with larger numbers of participants and with reading comprehension tasks.

According to the authors of the study, the key problem is the disruption of the process in which struggling with difficulties on your own allows you to develop skills, perseverance and creativity. Artificial intelligence, by offering immediate answers, eliminates this stage, so users become accustomed to quick solutions and lose the motivation to think independently. We built these systems to be helpful, but do they actually help us in a way that grows us? - Brian Christian from the University of California at Berkeley, one of the authors of the study, asks rhetorically. And he answers that, paradoxically, these systems often help us in ways that are unfavorable to us.

The results of these studies, which cannot be considered unexpected, raise particular concern in education and the academic community. Scientists point out that excessive reliance on AI may lead to shallow knowledge and loss of key competencies, such as critical thinking or teamwork. In the long term, this may threaten the quality of research and innovation. It's not just a matter of solving math problems or reading comprehension. This is a problem that affects all human knowledge and competences, from primary school students to world-class experts - emphasize the authors of the publication.

Experts suggest that the future should be based on the development of AI tools that will support and not replace human abilities. Instead of providing ready-made answers, artificial intelligence could act as an interactive teacher that guides the user through the problem-solving process, encouraging independent thinking. Researchers call for caution and caution when using new technologies, especially in education and science. Excessive automation may lead to the loss of what is most valuable: human curiosity, perseverance and the ability to creatively solve problems.