Artificial Intelligence for teachers according to UNESCO

License Creative Commons 

Antoni Hernández-Fernández

The Artificial Intelligence (AI) Competency Framework for Teachers, published by UNESCO in August 2024, addresses technological advancements in AI and their impact on education. What do teachers need to know about AI, and how can they apply it in the classroom?

UNESCO has released an insightful guide, the AI Competency Framework for Teachers, which is summarised and commented upon here. It is well worth reading and reflecting on. What does this framework refer to, and how does AI transform teaching? How does it contribute to the extensive body of work written in recent years?

Unlike previous technologies, UNESCO warns, AI could soon replace data-driven educational decision-making. This presents the risk that teachers may lose key competencies if they rely too heavily on AI. Therefore, the framework emphasises the importance of maintaining human agency and a people-centred mindset, ensuring that AI serves as a supportive tool rather than a substitute for teachers in performing their functions or making decisions. It is worth recalling the 2017 Barcelona Declaration for the responsible use of AI, which outlined six key principles:

1. Prudence: It is crucial to recognise the limitations and risks of AI before using it. This is the first principle, and one we often overlook.
2. Reliability: AI systems must undergo rigorous testing for safety, reliability, and dependability, particularly in domains such as medicine, education, and autonomous robots.
3. Responsibility: AI systems should be capable of explaining their decisions in a comprehensible manner. Individuals affected by AI decisions have the right to challenge them and receive clear explanations, with accountable individuals behind these tools.
4. Accountability: In this regard, it is essential to trace and hold accountable those responsible for AI systems. It is not acceptable to wash one’s hands of responsibility and simply blame the machine.
5. Restricted Autonomy: AI systems that operate in the real world must be subject to clear rules that limit their actions and ensure alignment with human values. This is particularly crucial in dangerous fields such as the development of autonomous weapons.
6. Human Role: Human intelligence remains crucial, especially in unforeseen situations for AI systems. AI must be supervised: synergy between AI and humans is key.

The use of AI raises numerous ethical challenges, such as data privacy, biases, and the potential for increased social inequalities. Its use in education is particularly high-risk, according to European regulations. Teachers must be aware of these risks and be trained in the responsible use of AI. Unlike earlier digital tools, generative AI, given a simple prompt, can produce outputs (texts, images, videos) stochastically, and it sometimes errs. Thus, it requires supervision. Teachers need to understand how AI functions to monitor it and critically assess its results, deciding whether or not to apply them in their work and determining whether the classroom activities it suggests are appropriate—can they be solved with the click of a button?

The 2024 UNESCO report proposes a human-centred approach, guided by ethical principles such as human development, equity, inclusion, and accountability in AI use. It is complemented by another, more extensive report specifically aimed at students, which is also worth revisiting and cross-referencing with the teachers’ report.

Teachers must be trained to have a fundamental understanding of AI, integrate it into their pedagogical specialisations, and promote social and cultural values, as well as its ethical and critical use, in a societal context where AI should not replace human interaction. Explainability will be one of the challenges in the coming years.

The use of AI in education should have been validated in terms of reliability, safety, and sustainability before its large-scale implementation. Teachers need ongoing support to adapt to these technological changes throughout their professional careers and must not be left to navigate these changes alone. Continuous training and policies that encourage lifelong professional learning are essential to this end, adapting curricula and assessment methods to the new possibilities that AI offers, with great caution, considering the high risks in education.

The structure of the AI competency framework for teachers proposed by UNESCO is presented as a two-dimensional matrix that includes five aspects which evolve across three progression levels (acquisition, deepening, and creation), forming fifteen distinct competency blocks (Table 1). By intersecting these three progression levels with the five competency areas, the framework defines fifteen blocks designed to support all teachers, from those with no knowledge of AI at all to those with a higher degree of competency and experience in AI.

Table 1. The AI competency framework high-level structure: aspects and progression levels, according to UNESCO (2024)

 https://unesdoc.unesco.org/ark:/48223/pf0000391104

First Dimension: The Five Competency Aspects

The competency aspects represent the key elements of knowledge, skills, values, and attitudes that teachers must develop to effectively and ethically integrate AI into the classroom. According to UNESCO, these aspects are complementary, interdependent, and synergistic. The five aspects are:

1. Human-centred mindset: Defines the critical values and attitudes necessary for interactions between humans and AI-based systems that teachers should promote.
2. AI Ethics: Establishes essential ethical principles, along with institutional laws, regulations and rules that teachers need to understand, apply and adapt.
3. AI Foundations and Applications: Specifies the transferable knowledge and skills teachers need to select, apply, and personalise AI tools for teaching and learning environments, fostering creativity.
4. AI Pedagogy: Involves effectively integrating AI into pedagogical methodologies and strategies, from course preparation to assessment.
5. AI for Professional Development: Details the competencies teachers must develop to use AI in their ongoing professional learning.

It is worth recalling the distinction between what is legal and what is ethical. UNESCO’s report tends to blend both elements at times: compulsory education should equip citizens to understand AI laws and data protection, as well as their rights and responsibilities, including what is legal and what is not. Then, there is ethics. For example, while it may be legal to use ChatGPT, it could be more ethical to use other large language models, such as HuggingChat, which are more transparent and don’t retain users’ data without consent.

Second Dimension: Progression Levels

UNESCO’s framework acknowledges that professional development related to AI is a complex, context-dependent process, not necessarily linear or hierarchical. Teachers’ interests and needs may vary depending on the educational levels they teach and their specialisations. However, the framework outlines three general progression levels to guide teachers’ development:

1. Acquire: Establishes the essential set of AI competencies all teachers need to evaluate, select, and appropriately use AI tools.
2. Deepen: Defines intermediate competencies required to design pedagogical strategies and teaching materials that integrate AI in the classroom.
3. Create: Establishes advanced competencies for creatively configuring AI systems and innovatively using them in education.

It is commendable that the complexity of teachers’ professional development is recognised, although it is worth noting that universal solutions for teacher training may not always be effective. Nonetheless, UNESCO’s effort in this document is notable, and I would recommend reading its fifty pages, though one might disagree with certain approaches.

In my opinion, the success of AI integration in education depends on basic factors such as those outlined in the Barcelona Declaration, access to digital infrastructure, stricter regulations on data privacy and security (which are already urgent considering the educational platforms that hold our students’ data), and educational policies that provide real support and promote professional development. This must include constant, classroom-relevant training opportunities, alongside employment incentives.

Furthermore, there are social aspects, such as the varying levels of trust from school leadership in the numerous AI tools available, or the potential for unequal workloads among teachers, which will affect implementation in schools: will more be expected in less time from those who are knowledgeable, while the same is demanded from those who are not? Isn’t this already the case with other technologies?

This coming December will mark 50 years since the famous Haifa conference, where Mario Bunge coined the term “technoethics”. In his manifesto Towards a Technoethics (Bunge, 2019:133), Bunge warned us: “Instruments are morally inert and socially not responsible. Hence, when acting as a tool, the scientist, engineer, or manager will refuse to take any blame”.

The same could be applied to teachers. Let us not become mere instruments; let us assume responsibility, be proactive, and show critical thinking in the face of what lies ahead.

References:

–  Bunge, M. (2019). Filosofia de la tecnologia. Barcelona: Edicions IEC-UPC.

Available at: https://upcommons.upc.edu/handle/2117/169030

Source: educational EVIDENCE

Rights: Creative Commons