Speech given by OECD Secretary-General Mathias Cormann to launch the PISA 2025 results on 8 September 2026.
Check against delivery.
Good morning, good afternoon and good evening to everyone joining us around the world.
Welcome to the global launch of PISA 2025 – the Programme for International Student Assessment (PISA) of the Organisation for Economic Co-operation and Development (OECD).
This is the biggest PISA ever.
More than 760 000 students took the test, across 91 countries and economies.
We don’t test every student. Like an opinion poll, we test a carefully chosen sample in each country.
So together, these 760 000 students represent around 33 million 15-year-olds in schools around the world.
Students were tested in science, the main focus this time, as well as in reading and mathematics.
For the first time, PISA also tested how well students learn and solve problems in the digital world.
So what did we find?
The picture is mixed.
In science, results held steady between 2022 and 2025.
That is welcome news. Science scores had been falling since 2009.
Science matters for every student, whatever they want to do in life.
It teaches them how the natural world works, how to test ideas, and how to solve real problems.
But holding steady is not enough.
We need to win back the ground lost over the past fifteen years.
Some countries have shown it can be done.
Since 2022, science scores went up in Costa Rica, Slovakia, Türkiye and the United Kingdom – and in 12 other countries and economies.
Over two decades, Türkiye has moved from below the OECD average in science to above it.
Cambodia gained 35 points in just three years – well over a year of learning.
The news in mathematics and reading however is bad.
Scores fell sharply, to the lowest levels PISA has ever recorded.
Across OECD countries, mathematics scores fell by 9 points between 2022 and 2025.
That is about half a school year of learning – and it comes on top of the record 15-point drop between 2018 and 2022.
Reading fell even further by 14 points.
That is the biggest drop between two PISA rounds we have ever seen.
Reading matters most of all, because it is the gateway to every other subject. If you cannot read well, you cannot learn well – in science, in history or in mathematics.
And the countries with the biggest falls in reading also had the biggest falls in mathematics and science.
Across the OECD, one in five students is now a low performer in all three subjects.
In 2022, it was one in six.
These students lack the basic skills they need for further study, for work and for everyday life – like understanding a simple text, or working out which product is better value at the supermarket.
This matters far beyond the classroom.
Weaker skills mean weaker productivity, less equal opportunity and a harder time adapting to a world being reshaped by technology.
These declines did not just start with the COVID-19 pandemic.
In many countries, results were already slipping well before 2020. Something deeper is going on.
In reading, more screen time and less reading for pleasure have gone hand in hand with weaker results.
Since 2022, the share of students spending more than an hour a day on social media has risen by almost six percentage points.
We are also seeing more "hasty reading" – students rushing through a text and giving a quick, wrong answer. The share of hasty readers has nearly doubled since 2018.
Digital tools can help students learn – or hold them back.
It all depends on how they are used.
Using devices for learning at school went with better science results – up to a point. Beyond about five hours a day, results start to fall.
And when devices are used for fun during class, results fall sharply.
Distraction is a real problem.
More than one in four students say their classmates are distracted by digital devices in most or every science lesson.
Family background still shapes results far too much.
Across OECD countries, students from well-off families scored 85 points higher in science than students from poorer families.
That is a gap of about four years of learning.
That gap did narrow between 2022 and 2025. But only because advantaged students did worse – not because disadvantaged students did better.
That is obviously not progress.
We want both to do better.
Real progress means lifting disadvantaged students up while keeping standards high for everyone.
Some countries have done exactly that.
Over the past decade, Türkiye, the United Arab Emirates, Montenegro, Qatar and Slovakia all cut their share of low performers and grew their share of top performers at the same time.
So better results and fairer results can go together.
How do we turn this around? Six priorities stand out.
First, set high and consistent expectations for every student.
Teach fewer things, in greater depth. It is better to teach a smaller set of core topics until students truly master them than to race through a long list.
And make sure the curriculum, the teaching and the assessment all pull in the same direction.
Second, give teachers what they need to do their best work.
Teachers need ongoing training – not just in what they teach, but in how they teach it.
The best teachers can mentor those who need more support.
And paperwork can be cut – including with the help of artificial intelligence (AI) – so teachers can spend more time with their students.
Third, get parents involved.
Something as simple as a parent asking "what did you do at school today?" goes with better science results and a more positive attitude to learning.
But it is happening less and less.
In 2022, 77% of students said their parents asked them about their school day at least once a week.
In 2025, it was 71%. The drop is biggest among boys and disadvantaged students.
Schools and governments need new and creative ways to bring parents back in.
Fourth, keep our best young scientists in science.
Across the 91 countries and economies, around 1.9 million students are top performers in science.
Yet almost half of them do not expect to be working in science, technology, engineering or mathematics (STEM) at age 30.
That is talent we cannot afford to lose.
Our economies need more scientific and technical skills, not fewer – especially as our populations age.
Students need science lessons that excite them, and better guidance about where science can take them.
Fifth, give the most help to the schools and students who need it most.
Across OECD countries, around four in ten students are in schools where the principal says a shortage of teachers is holding back learning.
Shortages of materials and support staff are getting worse too – especially in disadvantaged schools.
Funding, experienced teachers, support staff and learning materials should go first to the schools where students face the biggest barriers.
Sixth, use AI and digital tools in the classroom carefully, purposefully, and in moderation.
Students who used AI to write text for their assignments scored lower in science, on average, than students who did not.
AI can be useful – for feedback, or for personalised practice.
But students learn by wrestling with hard questions and working out the answers for themselves. AI should support that thinking, not replace it.
Schools must also teach students to question what AI tells them: check the evidence, compare it with reliable sources, and spot answers that are wrong or misleading.
At the same time, schools must prepare students for the growing opportunities in AI and computer science.
Our new "Learning in the Digital World" assessment gives us a first look at this.
The early results show we have work to do. Only 64% of students reached the level we consider proficient in computational problem solving.
We will publish the full results, and what they mean for policy, in May 2027.
In conclusion,
- PISA 2025 points to a clear way forward for our education systems:
- Focus on mastering the basics.
- Back our teachers.
- Bring parents in.
- Keep young talent in science and technology.
- Give the most help to the schools that need it most.
- Use technology to deepen learning, not to skip it.
The OECD will keep helping countries learn from each other, and turn this evidence into better policy and better practice.
Decline is not inevitable. The fastest-improving countries prove that.
But turning things around takes focus, discipline and action.
I now hand over to Andreas Schleicher, our Director for Education and Skills, who will take you through the results in detail.
Thank you.
Working with over 100 countries, the OECD is a global policy forum that promotes policies to preserve individual liberty and improve the economic and social well-being of people around the world.