The educational landscape is undergoing a significant transformation, with artificial intelligence playing an increasingly prominent role. However, recent data from the Programme for International Student Assessment (PISA) paints a concerning picture of global student performance. As educators, parents, and policymakers grapple with these challenges, the complex relationship between AI and learning comes into sharp focus.
In the early 2000s, the OECD initiated the PISA assessments to evaluate the skills of 15-year-olds in math, reading, and science. The latest findings indicate a worrying trend: scores in all three subjects have reached their lowest levels since the inception of the tests. This decline is particularly pronounced in reading scores which have plummeted by the equivalent of more than a year of education between 2015 and 2025. In the United States, reading scores dropped by approximately 14 percentage points between 2022 and 2025.
The PISA report: Key findings and implications
The PISA report, based on a sample of over 760,000 students from 91 countries, highlights several factors contributing to this decline. The lingering effects of the Covid-19 pandemic and an attentional crisis that hampers the ability of both students and adults to engage with lengthy texts are cited as primary culprits. Additionally, the report delves into the impact of generative AI on education, a topic that has sparked intense debate among educators and policymakers.
The role of AI in education: A nuanced perspective
The PISA report reveals that 86 percent of high school students use AI at least occasionally, with roughly one in 10 using it daily. However, the impact of AI on learning is far from straightforward. The report indicates that students who never use AI generally outperform those who do. Yet, weekly users who leverage AI to enhance their learning rather than merely complete tasks, perform comparably to non-users. Moreover, frequent AI users who have been taught AI literacy skills score slightly higher.
This nuanced data complicates the policy discussion surrounding AI and education. While some applications of AI in education are promising, others raise concerns. For instance, New York City has implemented a one-year moratorium on student-facing generative AI tools in public elementary and middle schools. Schools chancellor Kamar Samuels argues that this pause will allow the district to evaluate which tools and use cases genuinely benefit students.
AI applications in education: Promising developments
Despite the concerns, there are several promising applications of AI in education. Tools that alleviate teachers from administrative tasks and educational chatbots designed to provide personalized tutoring are just a few examples. In upstate New York, a federally funded research center is developing AI tools tailored for children with learning and language disorders. One mother shared her experience with an AI reading app, crediting it with significantly improving her autistic son’s communication skills.
The evolving role of AI in education presents both challenges and opportunities. As educators and policymakers navigate this complex landscape, the need for careful evaluation and thoughtful implementation of AI tools becomes increasingly apparent. The PISA report serves as a crucial starting point for understanding the impact of AI on learning and guiding future policy decisions.
