Artificial intelligence (AI) can help children to solve tasks and get better results in school, but ultimately may hinder learning and lead to laziness and disinterest, according to the OrganiSation for Economic Cooperation and Development (OECD).
Sophisticated AI chatbots can improve pupils’ task performance but don't necessarily lead to learning progress, researchers with the collective of data-sharing countries said, citing new findings.
When it comes to AI's ability to help children, there is also a discrepancy between task performance and genuine learning.
"Offloading cognitive tasks to general-purpose chatbots creates risks of metacognitive laziness and disengagement that may deter skill acquisition in the long run," the researchers said. Pupils’ mental effort to turn answers into understanding could therefore decline.
Research shows that pupils with access to AI chatbots achieve better results on tasks, the Paris-based organisation said. However, that advantage disappeared and sometimes even reversed when access to the AI tool was withdrawn during exams.
When students depend too much on AI, the mental processes and effort that turn answers into understanding declines, the researchers said.
By contrast, AI tools developed for educational purposes tended to deliver lasting learning improvements, the OECD said. It added that these specialized AI tools could help inexperienced teachers improve teaching quality and thus raise pupils’ learning outcomes.
The organisation advised deploying AI applications in schools selectively and purposefully to enrich learning, not to replace cognitive effort or weaken the human relationships at the heart of education.
This suggests that one of the most effective uses is not to replace humans, but to scale expert pedagogy across a variable workforce, the OECD said.
It said learning and teaching should primarily aim to develop valuable human knowledge and skills such as independent thinking and fundamental abilities across all subjects.
A team of United States-based researchers last year found that using AI instead of finding information using web search tools leads to no more than skin-deep knowledge, with any learnings soon turning "passive."
"When individuals learn about a topic from (AI), they risk developing shallower knowledge than when they learn through standard web search, even when the core facts in the results are the same," the team said last November. – dpa
