Using an AI chatbot to solve a problem, even just for ten minutes, could make you less capable of thinking and solving problems on your own, according to new research. Participants who used an AI assistant were more likely to give up or perform poorly when the helper was gone.

The study suggests that while AI can boost productivity in the short term, it might be eroding fundamental problem-solving skills. Researchers from Carnegie Mellon, MIT, Oxford and UCLA conducted three experiments involving several hundred people each, observing a significant drop in performance without assistance.

Michiel Bakker, an assistant professor at MIT involved with the study, suggests that AI’s role should not disempower humans but rather support their learning. He advocates for AI tools that prioritize user education over quick solutions.

The findings are particularly concerning as persistence is key to acquiring new skills and predicting long-term learning capacity. Future AI tools may need to be designed in a way that encourages users to solve problems on their own, perhaps by acting more like human teachers.

Reflecting on his experience with an AI assistant, the author found himself in a situation where he needed to learn how to fix issues for himself rather than having them solved automatically. This raises questions about the long-term effects of relying too heavily on AI and suggests that a balance must be struck between efficiency and skill development.