Students who use AI tools extensively may struggle with knowledge retention, according to new research.
Brazilian social scientist Andre Barcaui looked at two groups of students, one using ChatGPT as a study aid and the other using more traditional methods, before giving them a surprise test after 45 days. He found that those who had depended on AI scored an average of 57.5 percent on a knowledge retention test, compared to an average of 68.5 percent for those who had studied traditionally.
This randomized controlled trial showed that unrestricted use of ChatGPT as a study aid can impair long-term knowledge retention, Barcaui said in the conclusion of his paper, ChatGPT as a cognitive crutch: Evidence from a randomized controlled trial on knowledge retention. “Students who learned without AI retained substantially more information after 45 days than those who used ChatGPT.” He pointed out that an 11 percent performance gap was a significant differential.
A survey of around 10,000 teachers by the British National Education Union found other deleterious effects of AI usage. It found that two-thirds of secondary-school teachers (66 per cent) thought pupils’ critical thinking has declined due to AI usage.
The research will dismay those AI enthusiasts who have been promoting AI as an essential tool in education, although some observers have noted that there’s a need to be circumspect as to how it’s being used.
Critics will point out knowledge retention is just one part of education and that in-depth knowledge of AI will be of more help in the workplace.