PISA 2025 Global Launch — Reading, Mathematics, Science, AI & Digital Learning
Citation dossier for the OECD's global launch of PISA 2025. It covers record-low OECD averages in reading and mathematics, stable science performance, inequality and teacher shortages, digital distraction, student AI use, cyberbullying, family support and the OECD's early evidence on how AI can help or undercut learning.
Original Longscribe citation dossier grounded in the official source. This page is a paraphrased research asset and timestamped claim index, not a verbatim transcript.
Featured speakers
- Mathias Cormann
- Andreas Schleicher
Transcribed and published by Longscribe. Editorial policy.
The 2025 cycle gives the OECD a split headline: science stops falling on average, but reading and mathematics deteriorate further. The launch treats this as a structural education problem rather than a one-off pandemic effect because several of the negative trends began before 2020. Source · 01:42
Digital life runs through the event. Device use can support learning when it is purposeful, yet distraction, social media, rushed reading and cyberbullying are all associated with weaker outcomes. AI is presented as another tool whose educational value depends on whether it increases cognitive effort or simply produces finished answers. Source · 04:47
The OECD's policy message is not only about technology. It repeatedly returns to teacher quality, family engagement, social background, resource allocation and whether high-performing students can translate scientific ability into STEM aspirations. Source · 24:21
Key facts
- Largest PISA cycle
- More than 760,000 students across 91 countries and economies took part, representing roughly 33 million 15-year-olds. Source · 00:53
- Reading and mathematics record lows
- The OECD reports new average lows in reading and mathematics, while science remained broadly stable between 2022 and 2025. Source · 01:42
- Low performance across all three subjects
- The launch says the OECD-wide share of students struggling in science, mathematics and reading rose to about one in five. Source · 03:19
- Digital distraction
- The OECD links entertainment use during lessons and high levels of classroom device distraction with weaker learning outcomes. Source · 04:47
- AI and learning
- Moderate AI use for learning can perform differently from answer-generating use; the OECD emphasizes AI literacy and students doing the intellectual work themselves. Source · 33:18
- Family support
- The presentation identifies simple signs of parental interest in schoolwork as strongly associated with learning outcomes and notes that reported family support declined since 2022. Source · 39:07
Science stabilizes while reading and mathematics fall
The launch separates a modestly better science story from a much worse reading and mathematics story. Science stops its previous slide on average, but the OECD says reading and mathematics reach new lows. Country examples show that improvement is possible, which is why the event frames the decline as a policy challenge rather than an unavoidable global trend.
Equity, scientific talent and resource allocation
Schleicher uses country and school comparisons to show that social background and institutional design interact. The discussion highlights top performers, STEM aspirations, teacher shortages and the fact that disadvantaged schools often report greater resource constraints. The policy aim is to raise weaker students without lowering expectations for stronger ones.
AI, digital learning and cognitive effort
The OECD distinguishes technology that helps students think from technology that removes the need to think. Frequent AI use is not automatically beneficial, while moderate learning-focused use can look different when students have AI literacy. The same section connects device exposure with distraction and cyberbullying, making digital policy a learning-environment issue as well as a technology issue.
Teachers, families and lessons from East Asia
Teacher support and family interest remain strong correlates of learning. In press questions, Schleicher says several East Asian systems use AI in the background for teachers or through education-specific products built around pedagogy. He contrasts that model with general-purpose tools that can improve task completion without producing learning.
Key timestamped claims
Each claim is an original Longscribe summary anchored to the moment(s) in the recording where it is supported; follow a timestamp to verify it in the source video.
00:53PISA 2025 is described as the largest edition yet, covering more than 760,000 students in 91 countries and economies and adding a first assessment of learning in the digital world.
01:42OECD leaders say science performance held broadly steady from 2022 to 2025 while mathematics and reading fell to their lowest OECD averages on record.
04:47The launch links heavy non-learning device use and classroom distraction with weaker outcomes, while noting that purposeful device use can help up to a point.
08:55Cormann says schools should direct resources toward disadvantaged schools and use AI carefully so it supports practice and feedback without replacing students' own thinking.
12:08Schleicher shows that top science performance is concentrated in East Asia but also stresses that national income explains only part of the variation in outcomes.
18:23The OECD argues that scientific literacy includes checking evidence and questioning plausible-sounding claims, a skill it presents as especially important when AI can produce confident but incorrect answers.
33:18The presentation finds that frequent student AI use is often associated with lower PISA performance, but moderate learning-oriented use can be neutral or positive when paired with AI literacy.
41:20In Q&A, Schleicher says stronger East Asian implementations often keep AI teacher-facing or use education-specific tools designed around pedagogy rather than general-purpose answer generation.
Source & methodology
Source: Launch of PISA 2025 Results from the official publisher, duration 50:22. Official OECD replay downloaded from OECD TV and transcribed with Groq Whisper large-v3-turbo. This page is an original Longscribe paraphrased citation dossier, not a verbatim transcript. Every public factual block is grounded to committed source segments in docs/ops/curated-transcript-sources/OECD-PISA-2025-10898.json. The companion coverage manifest accounts for every source segment as covered by the claim index or explicitly excluded. Speaker names are used only when supported by official metadata or spoken introductions; no voice identity is inferred.
Paste an accessible public YouTube video or Shorts URL: you do not need to download it to your device. LongScribe uses captions when available or transcribes reachable speech with AI. Three hours is 180 transcription minutes; signed-in Free includes 60 minutes across 20 recordings per month, subject to remaining allowance, per-job limits and source availability. Completed eligible account jobs can export TXT, SRT, DOCX or JSON; the anonymous demo is a separate flow with its own limits.