This article is part of Come Invent With Us: Perspectives On OEGlobal 2026, a series of viewpoints on OEGlobal 2026’s program tracks by the Open Education practitioners who make up the OEGlobal26 Host Organizations.
In this piece, Jan Gondol — Director of Technology at Open Education Global (OEGlobal), and an open source, open data, and open government advocate who was downloading MIT lectures on a dial-up modem the year OpenCourseWare launched — argues that AI is open education’s next great frontier, and that the case for openness is as urgent as it has ever been.
__________________________
Twenty-five years after MIT put its courses online, the frontier is AI.
Contemplation time. (Want the argument, not the memoir? Skip to “***”.)
In 2001, I was a college student and I remember when the news dropped. MIT uploaded their courses online, for everyone to learn. I remember my heart beating fast. I downloaded tiny videos night after night on a dial-up modem. And watched lectures from a university I admired.
And then, other top schools followed suit and shared THEIR stuff. I was literally laughing, like “this can’t be real! What the hell is going on?”
The top dogs realized they weren’t competing for the best teaching materials. They were competing for the best students, so putting their courses online was free advertising. Ballsy, right?
There was something in the air in 2001: that same year, Wikipedia was born. Creative Commons, too. And folks met in Budapest nearby to define what “open access” in science would be. All within a few months, a true annus mirabilis.
Back at my school, we had to buy the “skriptá”, self-published textbooks. Some teachers sold them directly as a side gig and to discourage sharing, the professors kept a record and gave the “non customers” an extra hard time during exams. Many would drop a small update each year, so we couldn’t reuse last year’s edition (either bought or passed on by older classmates) and guess what was always reliably on the test? Yep, those small updates. The teachers were the gatekeepers. They captured all the value. And forced everyone to buy directly from them.
A complete opposite of what MIT did!
***
Twenty-five years later, Open Education and Open Access have earned their place in academia. They aren’t ubiquitous, but they are recognized and understood. There is now a new frontier: AI.
Artificial Intelligence is both exciting and risky in new and important ways. AI stirs up emotions, which makes rational discussions challenging: it was trained on people’s intellectual output, often in shady or outright illegal ways. It somehow gets even worse: what will the datacenters do to the environment? Will AI collapse the economy? Will it lead to massive job loss and societal upheaval? Will hallucinated slop and deepfakes hurt individuals and destabilize democracies? Will cyber weapons unleash an armageddon? There is a lot of fear, uncertainty and doubt (FUD) around artificial intelligence.
Some LLMs (such as those behind ChatGPT, Gemini, and Claude) are proprietary: their developers keep the model itself private and typically provide access only through their own services. Open-weight models make the trained model files available for download, so they can be run locally or on your own servers, although their licences may still impose restrictions. Open-source models go further, providing the weights, code, and information needed to study, modify, and redistribute them.
Without a strong push for open models, we may end up in a world where access to powerful AI models from the Western world runs through only a handful of companies, easily controlled by the government, inviting abuse. That would be a terrible world to live in – the fight is already under way.
The concentration of power is dangerous, it corrupts, and makes life harder for everyone. It corrupts professors who charge for their books, it corrupts companies, institutions and governments. As AI is becoming a critical infrastructure, much like electricity and the internet, we need to ensure that it will be as decentralized as possible. To deliberately push against the creation of oligopolies. And ensure that open LLMs are legally protected and widely used.
People in the various open movements already know how essential openness is. The frontier has shifted, from MOOCs, textbooks and articles to technologies like AI. New challengers and gatekeepers are springing up, but openness is as precious as ever.
H/T to IslaHF for providing feedback on an earlier draft.
__________________________
Jan Gondol is Director of Technology at Open Education Global. He served as Slovakia’s Open Government Partnership point of contact from 2014-2015, where he pioneered the inclusion of open education in the national action plan and helped lay foundations for OER and open access policy. He has collaborated with Creative Commons, SPARC, and international partners to advance open education, open data, open access, and open source software globally. He is also co-founder Erudo, the Slovak STEAM (science, technology, engineering, arts and math) civic association, which runs PyCon Slovakia and supports hundreds of teachers in teaching technology creatively and openly.
OEGlobal 2026 — Come Invent With Us! takes place October 7–9, 2026, at MIT’s Samberg Conference Center in Cambridge, Massachusetts, in a hybrid format open to participants worldwide. The conference is co-hosted by Open Education Global (OEGlobal), MIT Open Learning, and the Massachusetts Open & Low-Cost Educational Resources Advisory Council (OLERAC), and marks the 25th anniversary of MIT OpenCourseWare. Learn more and register at conference.oeglobal.org.