What does Europe’s new school AI package actually promise—and what is still only planned?
The Commission proposes AI-literacy assessment, shared education infrastructure and stronger teacher preparation. The package is a policy programme, not evidence that learning or teacher shortages have improved.
The short version
- What happened
- The package proposes age-appropriate AI-literacy outcomes and assessment guidance, including a pilot AI Literacy Certificate for 15-year-olds in the 2027/28 school year.
- Who it affects
- Pupils may encounter explicit AI-literacy expectations and future assessment, but the content, accessibility and consequences of the certificate pilot remain undefined.
- What remains unproven
- This is an official policy package, not an independent evaluation or a report of measured learning outcomes.
Keep in mind
This is an official policy package, not an independent evaluation or a report of measured learning outcomes.
Several measures are proposed Council recommendations or future pilots and are not yet binding across member states.
The documents do not provide complete budgets, participation denominators, service specifications or evaluation protocols for the AI measures.
PISA underachievement and teacher-shortage statistics do not establish that AI caused the underlying problems.
EU-level coordination cannot by itself determine national curricula, staffing, pay, procurement or classroom practice.
Source analysis · 2 linked sources. Primary analysed source date: 5 November 2026. Earliest linked supporting source: 4 November 2026.
Read the full report · 8 min The evidence, methods, implications and limitations
The direct answer: Europe has set a programme, not demonstrated an education result
The European Commission's Education and Skills Package puts artificial-intelligence literacy, shared digital infrastructure and teacher preparation into one EU-level programme. Its clearest AI-specific proposal is to develop age-appropriate learning outcomes and assessment guidance, then pilot an AI Literacy Certificate for 15-year-old learners in the 2027/28 school year. It also links school technology to EDU-EDIC, a consortium requested by eight founding member states to develop shared services and safer data exchange.
Those are consequential policy choices because they could influence what millions of pupils are expected to understand about AI and which systems schools are encouraged to use. They are not results. The Commission has not shown that the certificate improves judgement, that shared infrastructure raises attainment, that teacher training reduces workload or that the package closes inequalities. Several elements are recommendations for the Council or future initiatives whose design, funding and delivery remain to be settled.[1][2]
What the package says pupils should gain
The package treats AI literacy as part of a wider basic-skills agenda covering reading, mathematics, science, digital competence and citizenship. The Commission says it will build on the EU–OECD AI Literacy Framework to specify what learners should know and be able to do at different ages. The proposed certificate is framed as a pilot rather than a compulsory qualification, and the factsheet gives the 2027/28 school year as the intended starting point.
That distinction matters. A certificate can encourage curriculum attention, but it can also reward easily tested vocabulary while missing whether pupils recognise manipulated content, challenge an automated recommendation or understand when a model is unreliable. The package does not yet publish a validated test, pass standard, accessibility plan or evidence that certification changes behaviour. A credible pilot would need to report who participates, how tasks are scored, subgroup results and whether knowledge transfers beyond the assessment.[1][2]
The distinction that matters
Policy commitment versus measured outcome
The package contains concrete next steps, but none yet demonstrates education improvement.
AI literacy
Develop outcomes and pilot a certificate
Planned for 2027/28
Infrastructure
Pool trusted AI and cloud services
EDU-EDIC established; performance unmeasured
Teachers
Recommend better careers and AI preparation
Depends on member-state adoption
Learning
Raise basic-skills attainment
No outcome yet reported
EDU-EDIC is the infrastructure promise
The European Digital Infrastructure Consortium on Education was established on 3 November at the request of eight member states. The Commission says it will pool expertise and resources to develop trusted AI and cloud services for schools, enable secure exchange of data between education institutions and help European education-technology companies operate across borders. It is also presented as a route to greater European technological sovereignty.
A shared structure could reduce duplicated procurement and make interoperability or privacy requirements easier to enforce. It could also concentrate decisions about student data and approved tools. The announcement does not provide service-level performance, a completed privacy impact assessment, procurement totals or evidence that the planned systems work in different languages and school contexts. 'Trusted' is therefore an intended property that must be demonstrated through governance, security testing, accessibility and public reporting.[1]
Teachers are treated as professionals, but delivery sits with national systems
The proposed EU Teachers Agenda asks member states to improve career conditions, initial education and continuing professional development while giving teachers greater agency over AI use. From 2027, a revised generation of Erasmus+ Teacher Academies is intended to support novice teachers and school leaders, including preparation for child-centred uses of AI, diverse classrooms, digital change and complex learning environments.
This framing is stronger than simply telling teachers to adopt tools. It recognises that safe classroom use depends on workload, judgement, training and collective decision-making. Yet the Commission cannot directly set most teachers' pay, staffing ratios or day-to-day conditions. The impact will depend on national budgets, collective bargaining, local leadership and whether professional development is protected time rather than an additional task. No denominator is supplied for how many teachers will receive training or by when.[1][2]
The package responds to poor indicators, but those indicators are not caused by AI
The Commission cites PISA 2025 figures showing that 33.9% of EU 15-year-olds underachieved in mathematics, 31.3% in reading and 25.9% in science. It also reports that 40% of students attend schools where teacher shortages constrain learning and that 64% of 16- to 24-year-olds use generative AI. These statistics establish scale and urgency, but they do not show that AI caused the decline in basic skills or that AI-centred measures will reverse it.
The package combines several problems with different mechanisms: staffing shortages, attainment gaps, online safety, digital sovereignty and vocational responsiveness. Evaluating it will require separate measures. Learning outcomes need comparable assessments; teacher policy needs vacancy, retention and workload data; digital services need reliability and privacy evidence; AI literacy needs observable capability rather than self-reported confidence. One favourable headline metric should not be allowed to stand in for the whole programme.[1][2]
What families, schools and education-technology suppliers should expect
Families should expect more explicit discussion of AI literacy and online wellbeing, including a proposed European Healthy Screens Pledge intended to help schools, parents and pupils agree approaches to screen time, safety and balance between online and offline life. Schools may be invited into European alliances and advisory networks that share practices. These are support mechanisms, not new rights to a particular device, service or curriculum.
Suppliers may see clearer routes to cross-border education markets through EDU-EDIC and future procurement guidance. That creates an opportunity for smaller European firms, but it also raises questions about evidence thresholds and vendor influence. Public buyers should require versioned performance, data-use limits, accessibility, incident reporting, exit arrangements and proof that an AI feature addresses a defined educational need. A European label or consortium connection should not replace product-level scrutiny.[1][2]
What would change the assessment
The assessment would strengthen if the Council adopts clear recommendations, member states publish funded implementation plans and the Commission reports denominators, timelines and outcome measures. The AI-literacy pilot should be independently evaluated across languages, disabilities and socioeconomic groups. EDU-EDIC services should publish security, privacy, interoperability and classroom-utility evidence before widespread procurement.
It would weaken if the package remains a collection of future dates, if teacher development is unfunded, if certification becomes a proxy for meaningful literacy or if shared infrastructure locks schools into systems without transparent evaluation. For now, the package is a significant direction-setting intervention: it makes AI literacy and teacher agency explicit EU priorities, while leaving the proof of learning, safety and workforce benefit for implementation to supply.[1][2]