This website has been translated for your convenience, but no automatic translation is perfect. The official text is the English version of the website. If any questions arise related to the accuracy of the translation, please refer to the English version.

TUAC calls for education investment as PISA 2025 reveals deepening crisis

The latest PISA results expose a mounting crisis in education across the OECD, with student performance at its lowest recorded levels in science, reading and mathematics. TUAC calls on governments to invest in quality public education, reject the false promise that AI tools can substitute for ...

Share article:

Continue reading

pexels-yankrukov-8199151

The latest PISA results expose a mounting crisis in education across the OECD, with student performance at its lowest recorded levels in science, reading and mathematics. TUAC calls on governments to invest in quality public education, reject the false promise that AI tools can substitute for properly resourced schools, and tackle the economic inequality that leaves millions of young people without the support they need to learn. 

Published on 8 September, the 2025 assessment tested over 760,000 fifteen-year-olds across 91 countries and economies – the broadest cohort in the programme’s nine cycles since 2000. The scale of decline it documents is alarming: between 2015 and 2025, reading scores dropped by 28 points across the OECD and mathematics fell by 22 points, while science saw modest declines. The decline cuts deepest where it matters most: the ability to evaluate information, synthesise across sources and apply critical judgement – the very skills young people need in an AI-driven world. 

TUAC challenges the growing narrative that AI can rescue struggling education systems. Fifteen-year-olds who turn to AI chatbots for tasks like summarising readings, producing drafts or carrying out initial research score around 20 points lower in science than those who do not – a deficit roughly equivalent to one year of schooling. While students who combine regular AI use with school-based AI literacy training tend to score slightly higher, such opportunities are disproportionately available to socio-economically advantaged learners, reinforcing existing divides. 

Trade unions argue that these results point to a systemic failure to address inequality in education. In science, the performance gap between advantaged and disadvantaged students reached 85 points, and although it narrowed between 2022 and 2025, the shift largely came from declining performance at the top rather than progress for those who need it most. Support from families also fell compared to 2022, with the decline most pronounced among boys and disadvantaged learners – reflecting the wider toll that economic insecurity takes on households. 

Teacher shortages compound the problem. Four in ten students attend schools where principals report that a lack of teaching staff hinders instruction, and these shortages hit disadvantaged and public schools hardest. 

Governments must not outsource their responsibility for education to technology companies. AI is increasingly being pushed as a solution to the challenges facing education systems, but PISA shows that it can in fact hinder learning. AI use in education must be teacher-led and directed to specific pedagogical needs. Policymakers must invest in teachers, strengthen public education and confront the inequality that is undermining a generation's ability to learn.

— Veronica Nilsson, TUAC General Secretary