In May 2025, the European Commission and the OECD launched the draft AI Literacy Framework, with support from Code.org and leading international experts. The publication marked a key milestone in shaping how artificial intelligence is taught in primary and secondary schools worldwide.
Since the launch, the partnering organizations behind the framework have led a global consultation process, engaging teachers, education leaders, policymakers, learning designers, and other voices from across the education ecosystem to provide feedback on the draft.
More than 1,200 people participated in over 50 virtual and in-person consultations, while 800+ individuals from 97 countries across six continents completed the online survey. Stakeholders worldwide also led their own consultations with primary and secondary schools, universities, and organizations.
The framework’s partners also co-hosted several feedback sessions, including an event at the Digital Education Stakeholder Forum, a major European event on digital education and training, as well as a dedicated webinar for the community of the Digital Education Hub.
What We Learned
The draft framework is resonating with the global education community. Survey findings show that 84% of respondents believe the framework addresses a real need in education, while 81% indicate they are likely or very likely to adopt it in their work.
Notably, 79% agree that the framework successfully connects AI’s technical features with its societal implications. Among them, a teacher remarked:
“The AILit Framework balances technical, ethical, and pedagogical dimensions of AI literacy. Its structure makes it accessible for educators and policymakers, and the learner personas and classroom scenarios bring the competences to life.”
Several themes emerged consistently across consultations and survey responses:
- Reflection: Stakeholders emphasized the importance of metacognition and self-reflection when interacting with AI. This includes more emphasis on the need for students to recognize manipulation, regulate their emotions, and set boundaries when using AI.
- Agency: There was a strong call to position students and educators as shapers and decision-makers of AI, not just consumers of it. Hands-on learning and a focus on equity and social justice stood out as key priorities.
- User Support: Reviewers emphasized the need for practical resources to support classroom implementation. They also recommended using developmental progressions to make the framework more accessible for users.
- Healthy Skepticism: Stakeholders called for the framework to present a more cautious and critical stance toward AI, equipping students to critically examine how AI systems are designed and deployed.
- Environmental Impact: Many advocated for a stronger focus on AI’s environmental impact, encouraging students to consider the environmental costs of AI and make informed decisions accordingly
Implementing Feedback
Insights from the global consultation are shaping the next phase of the framework’s development. In September, the AI Literacy Framework Expert Group met at OECD headquarters in Paris to review the feedback and explore potential revisions.
The Expert Group reflected on the challenges of building a shared vision for AI literacy that serves diverse stakeholders. Pati Ruiz, Senior Director of Edtech and Emerging Technologies at Digital Promise, noted:
”What struck me most was the delicate balance we’re trying to achieve: creating a framework robust enough to withstand the rapid pace of innovation, yet flexible enough to adapt to diverse educational contexts and cultural values.”
Romina Cachia, Scientific Team Leader at the Joint Research Centre of the European Commission, added:
“Defining what truly constitutes AI Literacy for students isn’t easy — and that’s exactly what makes it exciting. The complexity of the conversations and discussions is a clear sign that we’re building something innovative and necessary.”
Analysis and revisions to the draft framework will continue through 2025. The final version will be published in 2026, alongside a limited set of AI literacy exemplars to support the implementation of the framework. We’ll continue sharing updates as the framework evolves.
