What was announced?
On 22 September 2026, the European Commission presented two reports on generative AI and digital education. One reviews evidence on AI in learning and teaching; the other examines digital education across European school systems. The Commission’s announcement is the source of the findings summarised here, not a study conducted by INNOQUEST.
The distinction that matters is simple: completing an assignment more successfully with AI does not automatically mean a learner understands more. The evidence review also raises concerns about over-reliance, while recognising that appropriately designed tools can support learning when students remain actively engaged.
What does the evidence say?
According to the Commission’s summary, Eurostat data cited in the paper indicate that four in ten EU residents aged 16–24 used generative AI for formal education in 2025. Separately, 29.7% of EU teachers surveyed in TALIS 2024 had participated in AI-related professional development. These figures concern different populations and years; neither measures the effectiveness of a particular product.
The Eurydice report describes uneven access to infrastructure and advanced tools. It also identifies gaps in teacher preparation and in the evaluation of digital education. Introducing another application is therefore not, by itself, evidence of a successful digital transformation.
Our interpretation for education projects
For an Erasmus+ cooperation proposal or a relevant Horizon Europe research topic, these findings support a concrete design question: what will participants be able to do without AI after practising with it? A simulation could let learners make a decision, explain their reasoning and repeat a task under changed conditions. Whether it improves learning remains a question to test, not a result to promise.
- Define a learning outcome and a transfer task before selecting an AI model.
- Include teacher preparation, accessibility and a non-AI participation route in the pilot plan.
- Compare outcomes with an appropriate baseline and report limitations, including who could not participate.
Where INNOQUEST fits
INNOQUEST’s proposed contribution is to co-develop decision-based exercises and pilot interfaces with education partners. Our AI-assisted training components remain in development. We do not claim that these reports validate our products or that any specific proposal is eligible for funding.
For the underlying method, read AI and simulation: learning through decisions. Our AI-Assisted Training note explains current readiness.
Source and dates
Source announcement: . INNOQUEST commentary published: . This is an independent summary and interpretation, not an EU endorsement.