This is the opening section of my September 19, 2024 presentation, Why Open Education Will Become Generative AI Education. I’m pre-posting some of the presentation content due to the very active conversation the announcement of the presentation has created. Last week I posted the middle section of the presentation, How Generative AI Affects Open Educational Resources, in which I described how we need to move beyond narrow thinking about how generative AI impacts our work with traditional OER and begin thinking more broadly about the power generative OER.
The Primary Goal of the Open Education Movement
For over 25 years, the primary goal of the open education movement has been to increase access to educational opportunities. The primary strategy for accomplishing this goal has been to increase access to educational materials. And the primary tactic for implementing this strategy has been to create and share OER. More than 25 years later now, it can be hard to remember what an incredible innovation OER was in the late 1990s.
Public access to the internet had just started to scale. Although they might not have learned the ECON 101 vocabulary terms, people were beginning to understand that when physical resources are converted into digital resources, they become essentially non-rivalrous. Take, for example, a physical copy of the New York Times. If my wife is reading a section during breakfast I can’t be reading that same section at the same time. Or for an even more dated example, if the Take 6 CD is in her car I can’t listen to it in my car. These physical resources (a printed newspaper or CD) are rivalrous. Wikipedia defines rivalry this way:
In economics, a good is said to be rivalrous if its consumption by one consumer prevents simultaneous consumption by other consumers, or if consumption by one party reduces the ability of another party to consume it. A good is considered non-rivalrous if, for any level of production, the cost of providing it to a marginal (additional) individual is zero.
The digital counterparts of these resources, however – the New York Times website or the Take 6 album on Spotify – are essentially non-rivalrous, meaning that many people can read and listen to them at the same time. In the late 1990s, many of us were beginning to realize that this same transformation could be applied to educational resources like textbooks. If you made them digital, then educational resources inherited this magical, non-rivalrous property and millions of people could all be using the same resource at the same time. This would obviously revolutionize the task of providing access to educational resources to learners around the globe.
But we have to pause for a moment. Just because it is technologically possible to share resources with millions of people simultaneously via the internet doesn’t mean it’s legal to do so. During a brief period of time beginning in the late 1990s, apps like Napster and LimeWire made it possible to share and download a wide range of digital music files. Many people got their first real introduction to copyright law during this period as they were subjected to countless PSAs about copyright and news stories about people being sued for millions of dollars for sharing music online illegally.
In some ways this was an incredibly frustrating period. The internet had made it possible to share at scale, but how could we make it legal to share at scale? Thanks to our old pal the Berne Convention (sarcasm), each and every creative work is automatically copyrighted to the full extent of the law at the moment it is fixed in a tangible (including digital) form. And this happens regardless of whether you want copyright protection for the work or not. But what if instead of protecting your work you wanted to share it with others? There was no easy or obvious way to do that.
In 1998 I launched the OpenContent project and released the first open source-style license for educational materials and other content. The next several years saw others pick up on this same idea. In 1999 I collaborated with Eric Raymond on the Open Publication License. In 2000 the Free Software Foundation released the GNU Free Documentation License (GFDL). And in 2002, Creative Commons published its first licenses.
What the internet had made possible, open licenses made legal. We finally had a simple, straightforward way to share educational materials (and other creative works) at scale.
The Most Powerful Emerging Technology
At the beginning of the modern open education movement in the late 1990s we were entering an information age – a time when the most powerful emerging technology (the internet) could make and transmit copies of digital resources almost instantaneously and at relatively low cost (compared to other options like shipping physical resources around the world). In this context, the tactic of creating and sharing openly licensed educational materials online (so that copies could be made and distributed freely and legally) was the optimal way to increase access to educational resources for people around the world. This tactic has proven more and more effective as internet connectivity has gotten faster, cheaper, and more broadly available.
Many people have forgotten – or perhaps never knew – that we used to pay per minute to use the internet. We dialed into modem banks using our landlines and listened for that marvelous sound that meant we were online. Compared to today, the internet of the 1990s was slow and expensive. But it was the fact that our usage was metered (charged per minute) that was in many ways the biggest barrier to the internet fulfilling its potential. When flat-fee, high-speed broadband finally arrived, people’s usage patterns changed and the internet revolution began in earnest.
Generative AI today is like the internet in the 1990s. Inference (“inference” is what it’s called when generative AI creates a response to your prompt) is slow, expensive, and metered. Even if you pay for a monthly subscription to something like ChatGPT Pro, if you submit too many prompts in too short a period of time you’ll get blocked temporarily and have to wait before you can submit additional prompts. But thanks to rapid advances in a range of areas – from algorithms to hardware – the speed of inference is increasing and the price of inference is decreasing. Generative AI is rapidly approaching the point where it will be, in Sam Altman’s words, “too cheap to meter.” In other words, generative AI is following the pattern established by the internet and will become inexpensive and ubiquitous.
This observation leads me to believe that we are transitioning from an information age to a generative age. While the most powerful emerging technology of the information age (the internet) could make and transmit copies of existing resources on demand, the most powerful emerging technology of the generative age (generative AI) can create new resources on demand. It is impossible for us to fully comprehend the changes that will be caused (both in education and elsewhere) by this capability. And those changes become more and more dramatic as generative AI becomes inexpensive and ubiquitous.
Generative AI Facilitates Greater Access than Traditional OER
Because it can create new resources on demand, generative AI can provide access to dramatically more resources, on more topics, in more languages, with more examples, using more pedagogies, in more formats, etc., than the current “create traditional OER by hand through a bespoke process” approach can. When we connect this fact back to the primary goal of the open education movement, the implication becomes clear. If:
- the primary goal of the open education movement is to increase access to educational opportunities, and
- the primary strategy for accomplishing this goal is to increase access to educational resources, and
- generative AI can provide access to dramatically more resources than the current bespoke OER authoring process can, then
- the optimal tactic for accomplishing the open education movement’s primary goal is no longer creating and sharing traditional OER –
- the optimal tactic for accomplishing the open education movement’s primary goal is to use generative AI.