ChatGPT use may weaken independent problem-solving skills

A large-scale study involving 1,222 participants from institutions including the University of California, Berkeley, MIT, and Oxford University suggests that reliance on ChatGPT may weaken independent problem-solving abilities.

Researchers observed that while participants using AI initially achieved high accuracy in mathematical and reading comprehension tasks, their performance declined sharply once access to the chatbot was restricted. Those who worked without AI assistance from the start demonstrated more stable results and a higher persistence in completing tasks.

Lead researcher Brian Christian noted a paradox where technology designed to assist humans may inadvertently hinder the development of independent skills by removing ‘productive struggle’—the essential process of encountering difficulties and searching for solutions to build new knowledge.