There is a new peer-reviewed study that has shed light on the long-term cognitive effects of excessive use of artificial intelligence tools like ChatGPT, Claude, and Gemini among university students.
The study was published in Social Sciences and Humanities Open and examines the effect of free access to AI aid on the learning outcomes.
The sample size was 120 participants, split into two groups. One group was free to use ChatGPT as an aid to study and the other used only the traditional learning methods without the support of AI.
Both groups were later on given a surprise retention test 45 days after their learning sessions to determine how well they learned the material.
The findings revealed that there was a significant disparity in performance. The mean score of students who used AI was 57.5 per cent as opposed to 68.5 per cent obtained by the students who studied without AI.
Researchers hypothesise that the reason behind this difference is a result of decreased mental effort during AI-assisted learning.
This process is called cognitive offloading and it is when people use external resources to process information rather than using their own memory and reasoning to do so.
The paper brings out the notion of desirable difficulties whereby effortful thinking contributes significantly to the enhancement of long-term memory.
When AI tools can give instant explanations or solutions, students can have a quick grasp of concepts yet not retain them over time.
Notably, the study does not purport that AI is a permanent killer of intelligence or creativity.
Rather, it cautions that such unfiltered and unstructured applications can lead to dependency, making AI a cognitive crutch.
The authors suggest a middle ground, according to which, with the help of structured and limited use of AI, students can use technology to their advantage without any negative impact on critical thinking and deep learning.
Also read: AI dominates professional Dota 2 match
