The 43-page document was developed by the European Commission, the EU’s main executive body, as well as the international Organization for Economic Cooperation and Development, with support from Code.org and experts from seven nations including the United States. It outlines what technical knowledge, human skills and ethical considerations will be crucial for students in the age of AI. It also defines four domains of AI literacy and offers specific lesson suggestions.
Educators and other stakeholders are invited to provide feedback on the framework before the final version is released next year in 24 languages.
THREE KEY THEMES
The framework states that three key themes — technical knowledge, human skills and ethical considerations — emerged from the study of existing AI literacy frameworks, curricula analysis, literature reviews, expert interviews and stakeholder focus groups.
The document defines technical knowledge as understanding how AI and machine learning work, and states that “learners must develop a strong understanding of AI’s technical foundations, including its reliance on data, probabilities and inputs.”
The human skills educators must help students build in order to use AI successfully include critical and computational thinking skills, the latter of which “assume additional relevance beyond the computer science classroom as students encounter technological challenges in their diverse everyday contexts,” according to the framework.
For ethical considerations, the document calls on school systems to treat ethical evaluation as a core AI skill rather than a supplement to technical concepts. The framework says AI lessons should include discussions about the ethics of how AI training data is collected and classified, as well as how algorithmic outputs can “reinforce existing patterns of unfairness if not critically examined.”
THE FOUR DOMAINS OF AI LITERACY
The framework goes on to provide examples of how to teach such technical knowledge, human skills and ethical considerations across what it calls the four domains of AI literacy. It defines each of these domains and then breaks them down into at least a handful of key concepts, each of which comes with a specific lesson suggestion for both primary and secondary students.
- Engaging with AI. This entails knowing where and how to access the technology for content, information or recommendations — and how to evaluate AI outputs for accuracy and relevance. One of the key concepts listed for this domain is to teach students how predictive AI systems work.
The suggested lesson for teachers of elementary school kids is to have them count by twos, fives and 10s; talk about how humans predict sequences; then explore how AI generates content based on patterns. For secondary students, the lesson involves examining how social media algorithms can contribute to the spread of misinformation. - Creating with AI. To teach students about creating with AI, the framework recommends lessons on prompts and feedback, content ownership and attribution. A suggested lesson for primary students is to compare human and AI art, then discuss how artists express themselves while AI generates art by using patterns in existing data.
- Managing AI. Managing AI means being able to intentionally choose when and how to use AI to support human work. Lessons here focus on learning how to tell which tasks are appropriate for AI versus those that should be handled by a human. For example, the document recommends having students talk about the benefits and drawbacks of using AI for specific tasks, such as writing a birthday card.
- Designing AI. This means helping students “build the confidence and capacity to shape AI for human good by understanding the principles underpinning the design of AI from an early age.” In this domain, a suggested lesson for secondary students is to talk about how AI could recommend after-school activities based on interests, as well as what data would be needed, how the AI would make its recommendations and which parts of the program should involve human input.
Cathy Adams, professor of educational computing at the University of Alberta in Canada and one of the experts who helped develop the framework, said such activities can help students understand that they are not powerless in a world of AI.
“Students don’t need to be AI engineers to design AI,” Adams said in the document. “Even simple, age-appropriate explorations of how AI systems work can spark powerful learning — and help students see they can shape technology, not just be shaped by it.”
