Returning to ECTEL After 15 Years

This year I returned to the European Conference on Technology-Enhanced Learning (ECTEL) after a 15-year break. I joined the pre-conference workshops in Newcastle and the main programme which took place in Durham. It was an opportunity to reconnect with a community that has always sat at the intersection of computer science, pedagogy and the social sciences.

Two Decades of TEL

Across its 20-year trajectory, ECTEL has shifted from a technology-led mindset toward a clearer pedagogy-first, human-centred orientation. Early work focused on e-learning platforms, learning objects and personalisation. When I last attended in 2010, learning analytics dominated debates: how to turn data into pedagogical insight, a question that still matters. Since then, themes have cycled through open education, sharing practices, gamification, mobile learning and immersive technologies. 

Today, AI in education is firmly centre stage.

Keynotes: Mike Sharples and Dragan Gašević

Mike Sharples opened with a long view, tracing generative methods back to his PhD work in the 1970s, a helpful context that tempers today’s sense of novelty. His recent paper A systems approach to AI and education in a post-digital world argues that AI should be treated not as an add-on but as an intrinsic part of educational ecosystems.

He framed learning as a dialogic practice and highlighted that, increasingly, the “other” in the dialogue may be a language-using technology.

Opportunities he emphasised include:

• More personalised tutoring and faster feedback
• Expanded pedagogical possibilities and administrative support
• AI as a lifelong learning mentor and research assistant

He flagged risks, such as:

• Academic integrity issues
• Unequal access, bias and data governance concerns
• Environmental costs and “autopilot” over-reliance

His core message was unambiguous: fragmented adoption will not work. Education needs a systemic, ethical approach to AI, one that preserves human agency, judgement and integrity. And teachers remain indispensable, not only for empathy but for modelling ethical reasoning.

Dragan Gašević followed with a keynote exploring tensions in AI-supported learning, where it can help, and where it quietly undermines key learning processes.

He discussed four paradoxes:

• Self-regulated learning & metacognitive laziness: AI risks making learners over-reliant, reducing reflective control.
• Learning vs. Performance: Faster ≠ deeper. AI helps us be more efficient, but that does not mean we are learning.
• Expertise & vigilance: Novices require high vigilance. AI can paradoxically reintroduce vigilance even for experts.
• Assessment of skill: Designing tasks around evaluating AI outputs is tempting – think for example of a task where learners are asked to produce an essay with the help of AI and then criticize/ evaluate it. That is a nice use of AI, but learners need practice in producing and writing themselves.

The key takeaway of his talk, in my perspective, was that AI must be used with care to support, not erode, self-regulation and expertise.

“… while LLMs like ChatGPT offer an efficient way to reduce intrinsic and extraneous cognitive load, they may not always facilitate the deep learning necessary for complex decision-making tasks. Traditional search engines, by necessitating more active engagement, may promote a higher quality of learning, underscoring the need for educational practices that encourage critical engagement with diverse information sources.” (link)

Pre-Conference Workshops 

Before moving to Durham, the community gathered for workshops in Newcastle. I joined the Dialogue Lab on the Role of Teachers and AI in Education (TAICo-Dialogue), which focused on policy questions around the evolving role of teachers in schools, universities and training organisations.

An interesting contribution came from the UCL team, the Five Levels of Teacher - AI Teaming (Cukurova et al.) which are:

1. Transactional: AI executes teacher requests.
2. Situational: AI informs teacher judgement.
3. Operational: AI supports enacting teacher goals.
4. Praxical: Teacher and AI refine each other through feedback.
5. Synergistic: Teacher and AI co-create solutions neither could produce alone.

The discussion emphasised the importance of aiming for praxical and synergistic teaming, so that AI amplifies teacher expertise and autonomy rather than substituting for it.

Poster sessions 

ECTEL’s poster sessions and demos showcased several promising directions. All materials are available on the conference website: https://ea-tel.eu/ectel2025/posters. A few examples include:

• SIMBA – A platform for exploring self-regulated learning with LLMs, and for modelling reflective learning processes. (Ferrettini, Nascimento, Pérez-Sanagustin & Hilliger)
• PromptHive – A collaborative interface for prompt authoring that connects domain knowledge with prompt engineering and enables rapid, systematic iteration. (Pardos, Bhandari & Anastasopoulos)
• Automatic LLM-generated scenario-based lessons – A task-decomposition approach to generating interactive math tutoring lessons for novice instructors. (Lin, Rao, Zhao, Wang, Barany, Ocumpaugh, Baker & Koedinger)

 

A Community Leaning Toward Pedagogy-First AI

Overall, ECTEL 2025 placed AI at the centre of discussion without succumbing to hype. The dominant stance was one of measured adoption, favouring:

• Human–AI collaboration rather than automation
• Pedagogy-first design
• Sustained human oversight
• Critical, not passive, use of AI tools

Before closing, the organisers also shared a light-hearted, but timely reminder: critical AI literacy is now essential for educators and learners alike. If you want to test your own assumptions, here's a challenge: the Global AILiteracy Test (GLAT).