Beyond the PISA rankings
The release of PISA 2025 has returned educational performance to the headlines. The results warrant attention, but they also invite a broader conversation about the relationship between knowledge, intellectual adaptability and the purpose of education.
More than 760,000 students participated in the assessment, representing approximately 33 million 15-year-olds across 91 countries and economies. Average performance in reading, mathematics and science has fallen to its lowest level since PISA began in 2000.
The decline in reading is particularly striking. The OECD average has fallen from 501 points at its 2012 peak to 466 in 2025. Mathematics has declined from 502 to 469, while science has fallen from 506 to 486.
These findings rightly return attention to the foundations of learning. Close reading, mathematical reasoning and scientific understanding shape how people evaluate evidence, distinguish correlation from causation and make sense of the world around them.
PISA, however, is not concerned solely with whether students can reproduce what they have been taught. The OECD describes its purpose as assessing whether they can extrapolate from their knowledge, apply it to new situations and function across different contexts.
The assessment has evolved to reflect this broader understanding of preparedness. Recent cycles have examined collaborative problem solving, global competence and creative thinking. PISA 2025 also assessed Learning in the Digital World, with those results expected in 2027, while PISA 2029 will add media and artificial intelligence literacy.
The questions graduates will face rarely belong to one discipline. Artificial intelligence is often framed as a technological question, but its implications extend into law, ethics, language, psychology, employment and power. Climate change requires scientific understanding alongside economics, politics, geography and an appreciation of human behaviour.
The most consequential challenges cut across disciplines, even though knowledge is necessarily developed and taught through them.
This is where the interdisciplinary character of a liberal arts education has particular relevance. At its best, it combines disciplinary grounding with the intellectual range to move beyond it. Students encounter different methods of inquiry, standards of evidence and interpretations of human experience.
They learn that another discipline does not simply add information to a question. It may change the question itself.
Interdisciplinarity depends on more than curricular breadth. It requires students to make substantive connections across fields while respecting the expertise and methods particular to each.
For universities, these ambitions are already embedded in program and student learning outcomes. The more searching question is whether they are coherently developed across the curriculum and evident in what graduates can actually do.
PISA concerns secondary education, but the questions it raises about knowledge, application and intellectual independence do not end at the university door.
A liberal arts education prepares graduates for careers that may develop in directions they cannot yet foresee. By cultivating intellectual agility alongside disciplinary knowledge, it enables them to carry expertise into new settings and continue learning as circumstances change.
This capacity to pivot is not an alternative to specialisation. It is what allows expertise to remain useful when the setting, the question or the profession changes.
AI reinforces this case. It can expand access to knowledge, support new forms of inquiry and help institutions understand learning in more sophisticated ways. Its educational value will depend partly on whether students can frame worthwhile questions, evaluate what it produces and place its insights within a wider intellectual and human context.
AI may also offer institutions richer evidence of learning through timely formative feedback and a more longitudinal view of students’ progress. It can contribute evidence and insight, while academic judgement remains essential to interpreting what that evidence means.
The question is not how to protect liberal arts learning from AI. It is how to use AI to extend learning while preserving the intellectual work through which students develop their own voice and judgement.
The latest PISA results should renew attention to the foundations of learning. They should also remind us that education must prepare students to use knowledge beyond the context in which it was acquired: across disciplines, in unfamiliar circumstances and throughout careers that may not follow a straight line.
That has long been one of the distinctive strengths of a liberal arts education. At a time when knowledge is more accessible, technology more capable and professional lives less predictable, that strength has acquired renewed significance.
