This article discusses the UNESCO Framework on AI Competencies for Students in the context of the draft Scottish AI Curriculum Framework. It explains the commonalities between the two frameworks and the rationale for not adopting certain aspects of the UNESCO framework in Scotland.
The UNESCO Framework on AI Competencies for Students is notable for focusing on human-centred responsible AI and promoting the skills required for learners to critique the place of AI within society. The UNESCO framework is based on principles of fostering a critical approach to AI; prioritising human-centred interaction with AI; encouraging environmentally sustainable AI; promoting inclusivity; and building core AI competencies for life-long learning. The ethical principles which learners should understand and apply are: do no harm, proportionality, non-discrimination, sustainability, human determination in human-AI collaboration, transparency and explainability, safe and responsible use and ethics by design.
There is much to be celebrated in the UNESCO framework. The framework is broadly consistent with the values of the education system in Scotland. However, the Scottish AI Literacy Curriculum Framework has made different choices about the relative weighting of sub-areas within the overall framework. The rationale for this is documented below.
The roles of learners as users or developers of AI
The UNESCO framework emphasises that all learners should exit the school system with a sophisticated technical understanding of how AI works, and practical knowledge of how to build and evaluate AI systems. While the UNESCO framework acknowledges that there is also a requirement for optional courses to further develop specialist technical skills, a very high level of technical expertise is suggested for all learners. In contrast, although the UNESCO framework touches on the requirement to learn how to responsibly interact with AI tools there is relatively little on how students could use AI tools to effectively support their learning. In the Scottish AI Literacy Curriculum Draft, we have chosen to explicitly develop students’ skills in using AI tools to support their learning. AI Literacy is included in the Scottish framework but the advanced level of technical detail is not required for all students. There are several reasons for this.
- Valuing breadth and choice. As a matter of principle, we believe that breadth of learning is important in broad general education and students should have a choice about the subjects in which they specialise. Learning time in schools is precious. Building strong technical focus on AI for all learners would require time in an already crowded curriculum. The education system, schools, teachers and learners may have legitimate other priorities.
- Curriculum reform process. Even if as a country we decided to focus our educational efforts towards producing highly skilled AI developers, several practical barriers would stand in the way. The first is that it would require the redevelopment of computing and maths curricula in broad general education and the senior phase. While a curriculum review process is underway it is unlikely that stakeholders would agree to the inclusion of perquisite topics for AI at the expense of other longer-standing and valuable areas of maths and computing required in other areas of life. The phasing in of the new maths and computing competencies would be a lengthy process and care would be required to support students to gain the necessary prerequisite skills in the transition years.
- Teacher availability. Upskilling the teacher workforce about AI will be challenging. To teach the more advanced topics in the UNESCO framework to all learners, more teachers would need a strong background in Maths or Computing and associated pedagogical knowledge. Unfortunately Scotland, like many other countries, already has a shortage of qualified teachers in these areas. There is a long standing issue about attracting computing graduates into the teaching profession; graduates with AI skills are in demand in industry and would need considerable incentive to become teachers. There are some schools in Scotland which do not have specialist computing teachers. Introducing a highly technical AI curriculm would further exacerbate this inequity of provision.
- Student interest. In the UK, there has historically been a lack of interest in computing subjects among students, particularly among girls. One of the factors which puts students off selecting the subject is its perceived difficulty. The UNESCO framework is extremely ambitious about advanced content (to the point that it contains several topics which are currently taught at university level). Requiring unmotivated learners to study difficult technical topics without a strong foundation or adequate access to teachers (at least in the short term) could be problematic.
- Knowing how to learn with AI matters across all subjects. The UNESCO framework has some content on learning how to use AI responsibly, but not specifically to support learning. It is appropriate to be cautious about claims of the educational benefits of emerging forms of AI as there is a lack of research evidence for the new generation of GenAI tools. However, there is existing evidence about previous generations of AI tools such as intelligent tutoring systems which are effective, particularly for STEM subjects. Mollick (2024) argues that the GenAI tools we use today are the worst we will ever use. It seems likely that AI tools will be increasingly used to support learning. We believe that in the spirit of human-determined collaboration with AI (a UNESCO principle), students should learn how to be in control of such tools and how to evaluate which learning tools to use under which circumstances, with the option to not use AI. This skillset will require instruction, modelling and guidance from teachers. In our view, students should have the opportunity to educationally benefit from advances in AI technology as well as being required to study the technology. This places a responsibility for understanding about AI on teachers of all subjects, but to a less technical degree than in the UNESCO framework.
Progression and pace of learning
The UNESCO framework specifies a spiral curriculum in four areas, with the stages of understand, apply and create in each area (human-centred mindset, ethics of AI, AI techniques and applications, AI system design). It lists the competencies that students should have when they leave school in each area. However, it deliberately does not elaborate on how learning should progress to these exit competencies throughout the student’s school career. As AI literacy is a new topic, little is currently known about how to support conceptual development about AI at different ages and stages. UNESCO leaves this an exercise to each country which implements it, which is understandable, but it does call into question what the dependencies and prerequisite learning might look like and whether the exit competencies are feasible.