This is the first in what I hope will become a regular series of podcasts on AI in education.  It has to be said there is no shortage of blogs, web sites, academic papers and now even blloks on AI in Education. But AI Pioneers is focused on the use of AI for teaching and learning in Adult and Vocational Education and Training and in particular on the role of teachers and trainers. For me this means pedagogy. So that is the focus of this first issue, looking at a recent, radical paper from Illka Tuomi and a call to arms from the UNESCO Assistant Director-General for Education, Stefania Giannini.

I’m still experimenting with different processes, content and form for podcasts. For this one I collaborated with ChatGPT in producing a transcript. So here is what the AI and myself came up with. Hope you enjoy.

AI Pioneers Podcast (Issue 1) Transcript

Host: Welcome to “The AI Pioneers Podcast,” where we explore the latest developments and insights in the field of artificial intelligence and its impact on education. I’m your host, Graham Attwell and I aim to produce this podcast as a monthly series. It forms part of the AI Pioneers project which is funded by the Erasmus+ programme of the European Union.  In this first edition we delve into ideas about knowledge development and the future of education. This includes talking about a thought-provoking academic paper by Ilka Tuomi entitled “Beyond Mastery: Toward a Broader Understanding of AI in Education.” Tuomi, a renowned researcher in AI and education, reflects on the origins and evolution of education systems and the implications for the future of AI in education. Join me as we unravel the complex and intriguing argument presented in this paper. And we will also report on a recent paper by Generative AI and the Further of Education by the ‌UNESCO Assistant Director-General for Education,‌ Stefania Giannini, highlighting the opportunities and challenges of generative AI tools for learning.

Before that I’ll give you some thoughts on where we are with AI in Education. And in wrapping up this podcast I’ll give you a quick rundown on what the AI Pioneers project is doing.

[Transition Music]

Host: Where are we now with AI and education? It seems like the initial shock and horror about AI is wearing off and institutions at least in higher education are realizing the need to educate students about how to use AI. In Vocational Education and Training, the pattern seems to be a bit slower. For this sector it is a harder challenge, not only how to use AI themselves for teaching and learning, but how to teach apprentices and students about the use of AI in very different occupational areas. At an institutional level in VET the picture seems very uneven, where there are pioneers or forward-looking managers projects and experiments with AI are flourishing but in other organisations staff may be waiting for support before trying anything bold. And as usual there is a need for more knowledge sharing between different organisations at regional national and European level. As our work has shown in the past, teachers and trainers often embrace change – especially when the work in their occupational areas is changing – but lack the professional development and training they need for the fast-changing world of work.

Just a few words about the thorny issue of assessment and accreditation. Well, calls to ban AI seem to be petering out. And attempts to use AI to expose the use of AI are failing badly. This software, often built on top of previous anti plagiarism applications is not very good at detecting AI and tests on seven popular AI text detectors found that articles written by people who did not speak English as a first language were often wrongly flagged as AI-generated, a bias that could have a serious impact on students, academics and job applicants. There is a general move toward more authentic assessment which can build in. the use of AI as part of the assessment. And as Ilka Tuomi says in his paper Beyond Mastery featured in the next section of this podcast, a move towards competence based assessment may be linked to the changes in our understanding of the purposes and education and knowledge.

[Transition Music]

Host: In 1964, teaching machines and programmed instruction saw a significant breakthrough in scaling up their implementation. However, much of this technology had lost touch with the fundamental principles of learning. Technology at that time primarily focused on testing rather than teaching. It was during the cognitive revolution that the behaviorism underlying early teaching machines was debunked. But now, operant conditioning is making a comeback, powered by data-driven AI. Tuomi suggests that data-driven AI stands on the shoulders of Skinner and behaviorism, employing positive and negative reinforcement in learning processes.

Host: To better understand the implications for AI in education, it is crucial to address two fundamental questions. First, why do societies have educational systems and what do they aim to achieve? Second, what is learning, and how should we measure it? Tuomi argues that these questions have been inadequately addressed in AI in Education research. Drawing from Biesta’s work in 2015, Tuomi highlights that education is not merely about enabling people to learn, as learning can happen anywhere. Education is about guiding individuals to learn specific things, for specific reasons, through particular educational relationships.

Educational technology has historically been developed from an engineering perspective, with a focus on measuring learning outcomes, collecting evidence, and defining objectives. This approach prioritizes efficiency and models of education where the mastery of predetermined learning content plays a central role. Bloom’s model, for instance, aims for mastery of learning within a group based on predefined learning objectives. Despite recent expansions toward social and emotional skills to meet the demands of the 21st century knowledge society, the influence of personalized tutors, as envisioned by Bloom in 1984, still lingers within the AI in Education community.

Tuomi challenges the assumption that personalized feedback should guide learners toward the same learning outcomes. Instead, he suggests that the learning outcomes themselves should be personalized. This departure may seem radical, as educational qualification based on achievement has played a vital role in sorting and stratifying societies over the last century. However, the emergence of new infrastructures of knowledge, including data-driven AI and new credentialing systems, renders the Bloomian model of mastery learning as an artifact of the past.

The educational landscape is undergoing a shift toward competence-based models. Although there is no consensus on the precise definition of “competence,” it can be viewed as the capability to effectively accomplish tasks, combining epistemic and non-epistemic competence components. Traditionally, education has focused on the epistemic part, encompassing knowledge, domain-specific skills, and experience. However, the current educational policy shift emphasizes the importance of non-epistemic competence components, often referred to as “soft skills,” “transversal skills,” or “21st-century skills.”

Tuomi highlights the challenges in maximizing or optimizing non-epistemic competence components such as emotional skills, conscientiousness, or creativity. The Skinnerian model of learning, based on behaviorism, stands in contrast to Vygotsky’s model, which emphasizes qualitative changes driven by practical action and social support. For Vygotsky, learning involves the development of conceptual structures that enable new forms of thinking. It is not merely about acquiring knowledge incrementally but about fostering disruptive capabilities to think and act differently.

Tuomi suggests that the acquisition of “higher forms of thought” expands our understanding of the world and blurs the line between epistemology and ontology. In this context, the notion of a learner model becomes less obvious. Learning in these new realities requires exploration, rather than knowledge transfer or incremental additions to existing knowledge structures. This distinguishes human systems of knowledge, rooted in evolving social and technology-mediated practices, from data-driven AI, which represents contemporary outputs of this process.

The paper also touches on the interplay between tacit and explicit knowledge, highlighting the challenge of modeling knowledge with computers. Nonaka’s work in the 1990s suggested that new knowledge emerges from the articulation and internalization of socially shared tacit knowledge into explicit representations. This perspective emphasizes the dynamic and process-oriented nature of knowing, contrasting with the object-oriented view that represents knowledge as static and stackable.

Looking toward the future of AI in education, Tuomi proposes a shift away from automating and sequencing knowledge delivery. Instead, he advocates for technology that supports learning and development, aligning with concepts such as open learner models, metacognitive support, and self-regulated learning. The objective of technology design should not be limited to efficient mastery of predefined universal learning objectives. Instead, the focus should be on developing systems that enhance personal capabilities, supporting individual learning and well-being.

That concludes our exploration of Ilka Tuomi’s paper, “Beyond Mastery: Toward a Broader Understanding of AI in Education.” We’ve examined the evolution of education systems, the significance of personalized learning outcomes, and the shift from epistemic to non-epistemic competence components. Tuomi’s thought-provoking arguments challenge us to rethink our approaches to AI in education and consider the potential of technology to augment personal capabilities.

[Transition Music]

Host: Early in July, UNESCO published a new paper, Generative AI and the Further of Education by the ‌UNESCO Assistant Director-General for Education Stefania Giannini highlighting the opportunities and challenges of generative AI tools for learning.

In her paper Stefania  considers the implications of AI for knowledge systems.  “Machines that offer immediate, concise and seemingly definitive answers to knowledge questions can be helpful to learners, teachers, and others”, she says. “But the technology can also usher in a world where machine knowledge becomes dominant, and proprietary AI models are elevated to global, and perhaps even revered, sources of authority. These models will project certain worldviews and ways of knowing and background others.”

This may be even more so in a situation with just one or two AI models and platforms, some of them already exercising near monopolistic powers, who are able  :to assert even greater dominance over our interface with knowledge,” Stefania considers that “As AI technology continues to permeate our world, we must preserve and safeguard the diversity of our knowledge systems and develop AI technologies in ways that will protect and expand our rich knowledge commons. We cannot allow our varied systems for producing knowledge to atrophy, and we must guard against delinking knowledge creation form human beings.“

In terms of education system Stefanie believes “AI is forcing us to ask questions about the ‘known-world’ that we usually take as a starting point for education. Many of our old assumptions and norms, especially those concerning knowledge and learning, appear unlikely to sustain the ‘weight’ of this new technology. We can no longer just ask ‘How do we prepare for an AI world?’ We must go deeper: ‘What should a world with AI look like? What roles should this powerful technology play? On whose terms? Who decides?’ Education systems need to return agency to learners and remind young people that we remain at the helm of technology.

[Transition Music]

Host: Just some quick information about the AI Pioneers project. The project aims to explore, promote and evaluate the use of AI for Adult Education and Vocational education and training, both the use for teaching and training and the curriculum needs of working with AI. The project also aims to support the development of new initiatives and projects using AI as a necessary step towards the mainstreaming of AI n education in Europe. The project aims to explore, promote and evaluate the use of AI for Adult Education and Vocational education and training, both the use for teaching and training and the curriculum needs of working with AI.

The definition of a Pioneer is drawn from the DigCompEdu Competence Framework, developed by the European Joint Research Centre in Seville. Pioneers experiment with highly innovative and complex digital technologies and / or develop novel pedagogical approaches. They lead innovation and are a role model for teachers and trainers.

A consultation meeting was held online in June with almost 50 people attending and the project is planning to launch an online social network, using the Mastodon software. for AI pioneers in the next couple of months. Are you an AI pioneer. Would you like to become an AI Pioneer.  Visit the AI pioneers web site at AIpioneers.org

Thank you for joining us on this episode of “The AI in Education Podcast.” Stay tuned for more exciting discussions and insights. Until next time!