I was recently invited to the 17th annual International Conference on Education and New Learning Technologies (EDULEARN25) in Palma, Spain. Getting to visit beautiful Palma, Mallorca, speaking with 800+ educators from across the world… what could be more awesome?

At the conference I gave a keynote and conducted a workshop (details and video below). But first, I must take a moment to thank the organizers for inviting me to wonderfully designed conference—with talks, workshops, breakout sessions, topped off by a wonderful sightseeing tour. A special note of gratitude to Inma Tomas for her generosity, kindness, and flexibility, especially given flight cancellations and travel complications. It was also wonderful to finally meet Laura McBain, the other keynote speaker. As it turns out, I had written a chapter for a book titled My Favorite Failure co-edited by Laura and my friend and colleague Ron Beghetto. Interestingly a version of that chapter ended up becoming a key part of my TED-Ed talk on How to Design a School for the Future.

Readers of my blog would not be surprised by some of the themes and ideas I explored in my keynote (“Education in an age of Generative AI”) and in an interview I did with the media team later. Essentially, I focused being on being more critical, and intentional about how we can use AI in educational contexts, recognizing both its capabilities and limitations. I shared my go-to metaphor for understanding and working with these technologies – calling them “smart, drunk, biased but supremely confident interns.” Essentially, these are dialogic, epistemic technologies and can do things we never thought technology could do. At the same time, we must recognize that confidently “making thing up” (aka hallucinating) is their modus-operandi. It is not a bug; it is a feature. And finally, since they are trained on us, and our myriad digitized outputs, they display many of the negative attributes that we humans have.

I also shared numerous examples that demonstrated both the remarkable capabilities and concerning limitations of these technologies. From writing mathematical poetry to identifying solar eclipses: from generating incoherent scientific diagrams to clocks that all show 10:10. On the creative side, I shared some of my typographical experiments with Claude and the interactive science and math simulations I have built using this technology. In essence, arguing that AI can serve as a powerful creative partner, but only when you bring domain knowledge, critical judgment and a care for learning to the collaboration.

My workshop “Human Creativity to the Power of AI” took this exploration further by moving beyond theoretical discussions into hands-on experience and playful discovery. With over 80 participants we had a lot of fun, understanding the pleasures and pitfalls of working with this strange technology.

In my interview and keynote I also spoke to some of the deeper long-term concerns I have shared on this website. At some level, the bigger challenge isn’t the AI itself, but the corporations behind it who prioritize engagement over user wellbeing, much like we’ve witnessed with social media. Finally, I made the case that as academics, we play a critical role as neutral third parties who can present both the positives and risks of this technology.

My keynote and the interview are embedded below.