Are OE, AI and commercialization compatible?

Hello OEG Community. Cathy Casserly here - reaching out to introduce myself and ask about the compatibility of OE, AI and commercialization.

For my introduction - my roots in Open Education extend from its earliest days when I was Program Officer at the Hewlett Foundation launching the initial portfolio of work for the field. MIT OCW was groundbreaking and there was global interest to replicate the model and learn jointly.

And so OEG (then OE Consortium) was launched. I had the privilege to serve as an early board member, and after some gap, currently serving again with an outstanding set of board colleagues. After Hewlett I was CEO of Creative Commons and now have my own consulting firm focused on strategic engagements and executive leadership coaching.

For the question about AI and commercialization. First, let me acknowledge that commercialization can be a sensitive word in the OE community. From my perspective, money is a pragmatic reality. Nonprofits need funds to operate, higher education institutions need fee paying students and often government support. This reality, however, doesn’t need to compromise the fundamental ethos of OE - of the open principles we all believe in.

Could it be possible for the OE community to commercialize OE for the benefit of the OE community, rather than await the threat of the corporate sector acting independently?

Could you please chime in with your reactions and thoughts? There are many conferences and discussions on AI and openness. I am specifically wondering how the community could potentially benefit from all of its incredible work, vis-a-vis a revenue stream, to sustain the community of work, as learning evolves with openness and AI.

My appreciation in advance for sharing your knowledge.


My experience with open education began in 1996. Since then, I’ve worked for mostly nonprofits but also for enough for-profits to get a fairly good understanding how the revenue streams flow. Here’s what I’ve noticed - your work either adds to the stream or it diverts revenue from the stream.

The nonprofit I founded seven years ago is currently leading a team that is building a tool to add to the stream. It is briefly summarized [here]. ( We’ll have more info soon. Let me know if you want to get involved.


Hello Cathy, thank you for posting here with your question. I certainly do not have all the answers, but as a recently redundant employee, I am thinking about consulting, starting my own non-profit, and how AI can benefit my open work and the work of others. In my values world, open practice should never be unpaid practice, and so it is critical to think about revenue generation in the context of open work. It’s challenging to think about it ethically, a fine balance. In terms of AI and the corporate sector, that horse is well out of the barn and the coming storm of snake oil promises for AI enhanced tech solutions is going to be a nightmare to navigate and protect against for those that work at education institutions. For me the process of open creation, adaptation, editing, etc. should be paid, the output should be fully open and grounded in FOSS platforms for absolute ease of access and no cost for people that want to learn. If I can invent some tools (and I may be able to) that use AI transparently and ethically that can enhance and make the process of OER creation more efficient and high quality, I’m 100% in. I would promote such a product and sell it to others that could benefit so that their work in open could be more productive and enhance their paid options for the work they do. I would even be on board with selling products for student use (at the lowest possible cost with the highest pedagogic value), but I would be extremely careful about it all. Just some of my current thoughts, I hope they are helpful! If anyone is looking for an AI learning and product development buddy, I want to learn much more about AI’s potential and applications and apply what I learn to my open work and the work of others!


Hear, hear, @Jenni! (And thanks, @cathcass for raising this important topic.) Sorry to hear about your recent status as a redundant employee, Jenni. Sounds like you have the right spirit of optimism in the face of such difficulty. Ensuring that we are not growing the field on the backs of exploited (often free) labor has been a soapbox of mine for some time now. And now I actually am a consultant, with work in Open Education (and AI as well) one of the areas of innovation I support. Some of it pays (thanks to those who may read this who have provided those opportunities!!!). But some of it doesn’t, and I have considered that work important to continue to be engaged in the field and contribute to its growth. But much of that uncompensated work isn’t converting to paid “gigs,” as I had hoped, and that isn’t good for the bottom line of my consulting business, I know – just isn’t entirely sustainable. I am not sure what the solutions are in my particular case, so I will follow this thread with great interest, hoping for ideas from minds more brilliant than my own.

Incidentally, this was the output for the prompt, using the image generator Craiyon, “illustration of optimism about the future”:


1 Like

Open becomes NOT Open when you sell it at any cost. You can’t sell OER. period. full stop. Now, to be sure, Jenni, you didn’t say you would be selling OER. But, if you’re selling things, materials, you’re not creating OER.

You can get paid for creating OER, and you can get paid for the service of implementing the OER, but you can’t sell OER at any cost. You can and should be paid for implementing OER in educational programs.

The fuzzy area that Cathy didn’t talk about is creating OER and licensing it CC BY and then cutting deals with proprietary vendors who sell it and kick back some money for ‘sustainability’ or some other fuzzy concept. This practice, as in the case of Illustrative Mathematics, in my opinion, is why OER is not able to gain traction in public education.

Education is not a market place. If a private school wants to buy things, fine. But government owned and operated public education should be only paying for the creation and service of implementing CC BY NC ‘products.’

Public schools should also go back to low bid contracts for non-educational materials.

1 Like

I applaud you @cathcass for launching a perhaps contentious topic that has/will generate differing viewpoints. That’s what good open discussion ought to be.

I am hopeful we might hear in this conversation from @paulstacey who has been looking broadly and eeply at the crossroads of AI and education, but also co-authored a few years ago, what ran likely against the conventions of Open People, the book Made With Creative Commons as a case study look at businesses and organizations navigating this space of entities earning money while at the same time publishing open “stuff”.

The word commercialization implies a motive of operating for the purpose of profit or at least “bringing something to market”. The business-minded CEO of the first educational non-profit organization always said that running a not for profit still meant applying business principles because (as Judith suggests above) the staff had mortgages and food to buy.

For this discussion, I do not believe Cathy was suggesting commercializing OERs or open education but to develop something that might provide the benefits or uses of AI that is growing (commercially) at wild rates of expansion. If we are talking about generative AI systems, there’s a reason why there are only a few source players, being the immense cost ($ and environmental)… I am a bit doubtful how even a large collaboration of educators could run that kind of system independent of the Big ___ (is it 4? 5?).

Every time I explore a new service mentioned, an early look on the web site menu is “pricing” – nearly all have one of those subscription pricing tables. And my hunch is 99% of them have developed software/systems built atop the APIs from the commercial providers.

At the same time the term “AI” is much more encompassing than the chat based prompt machines that get much attention. There are new developments happening too for means to put the technology to use (I am inferring) without relying on the Big AI Iron, e.g. Apple’s release of open source code for running LLMs on desktop machines.

Or maybe as a straddling the line small example is MacWhisper an app for AI transcription of audio, available in a well functioning free version and a more powerful $ version. It is built on top of an open source version of [whisper.ccpp(GitHub - ggerganov/whisper.cpp: Port of OpenAI's Whisper model in C/C++), so open seems to be at work out on the long tail of AI.

But to turn it back, what would some kind of education servicing AI provider do, look like?


What do we mean by “compatible” here? I don’t see specifics about what AI is being imagined to offer the open education community, which would be necessary to make informed statements about compatibility (with “open values”, with ethical expectations, etc).

Also, who is asking the question? Is it venture capital firms? Foundations with shifting interests? Educational institutions? Individuals? Not all opportunities are equally available to everyone, and so knowing who plans to exploit the responses is important too.

I’ve written before about what I refer to as “the turn” where for for-profit companies trying to find sustainable models for working with OER resort to enclosure, introduce scarcity, sell out to for-publishers, etc because the traditional capitalist tactics that produce revenue streams are seen as the only way. So, there is a history of business interests being prioritized over community values and long-term goals.

Made With Creative Commons talks through many varying business models built around openly licensed content. To my understanding the book was written and published prior to the late rise in AI hype, and may not have touched on OER+AI much if at all.

This topic brings to mind technological solutionism, wherein AI is being proposed as the solution to a problem without clear indications of how it will solve the problem.

Does AI have the potential to accelerate or improve the production, discovery, use, or uptake of OER broadly? Perhaps. But whether revenue streams benefiting the open education community can be created alongside these yet-proven innovations is an entirely different question.

Thanks @danmcquire @Jenni @JudithSebesta @cogdog @billymeinke for jumping in with your thoughts. The question was truly meant as a prompt to consider how OER could be harnessed for the good of the OER community given the evolving world of AI and learning.

I agree with comments shared - open practice should be renumerated, as should developers of OER. And at the core of the mission of OER is that learners should always have free access without obstacles. Nothing changes that.

@danmcquire I agree, nonprofits and for profits are designed with very different missions, cultures and bottom lines. I do think some B corps are beginning to blur the lines. And of course, governments funding OE fully is the sustainable end game for our community, aka the holy grail.

@Jenni raises a great example of a potential tool that would that could accelerate the work of OER creation, at higher quality. Interesting thought. Jenni - sending out good vibrations to you and hope you are able to find the next great position soon.

@JudithSebesta I don’t think there are any easy ways of navigating this space - sounds like you have found some gems along the way - just need a few more. And I do like the output illustration - beautiful colors. With you, I hope for an interesting discussion.

@cogdog @billymeinke thanks for sending the links for further reading. And yes - cost is a huge issue to building AI systems.

The concern, my concern, is that some corporate entity, without the values of the OER community, will leverage OER for its financial gain. Here I am not considering working with proprietary vendors - seems we would not in control if we did that. Trying to think differently…

Billy ended with a great question that I will reiterate here:

Does AI have the potential to accelerate or improve the production, discovery, use, or uptake of OER broadly? Perhaps. But whether revenue streams benefiting the open education community can be created alongside these yet-proven innovations is an entirely different question.

To me, this is the question to noodle on…


Hi @cathcass

A significant conversion.

Article 23 of the Universal Declaration of Human Rights states that:

Everyone who works has the right to just and favourable remuneration ensuring for [themselves] and [their families] an existence worthy of human dignity, and supplemented, if necessary, by other means of social protection.

From my perspective - I have no problems with anyone earning a living from OER related activities derived from openly licensed resources. As an open educator, I do not wish to deny anyone the right to earn a living from my creative outputs.

My bigger concern with AI is the commercial platformisation of access to the potential of these technologies for open education. Those who could benefit most, will not be able to afford to use these technologies for their own benefit :-(.


“Those who could benefit most, will not be able to afford to use these technologies for their own benefit.” Well stated. Or be able to contribute to the development of the technologies as well as the policy and laws enacted around them. I find Kate Crawford’s book Atlas of AI helpful to inform such discussions. In it, she argues, "AI is not simply a technology; it is also a registry of power.” As such, it is crucial to ensure a balance of power – and to mitigate adverse effects when that power is wielded irresponsibly – among those affected by AI.

1 Like

Btw, I am co-developing and facilitating a course this spring, Navigating the Future: Open Education with Generative AI, for those creating ZTC pathway programs at California community colleges as a part of the statewide ZTC Acceleration Grants. Eventually the plan is to add the course to Canvas Commons, so stay tuned.

I rather enjoy the critical reviews of AI papers and articles in the Mystery AI Hype Theater 3000 podcast by linguist Emily M. Bender (of Stochastic Parrot fame) and sociologist Alex Hanna.

Just halfway through Episode 17

which opened with some critical remarks on University of Michigan’s system wide rollout of AI, as posted in ITS debuts custom artificial intelligence services across U‑M – really worth listening to.

1 Like

Hi @cathcass , and thanks for the nudge @cogdog. I’ve been a bit heads down developing a virtual keynote talk on “AI, ELearning and Open Education” for an E-learning and Open Education international conference in Taiwan this coming week. It’s been interesting developing that talk as it has forced me to look at what AI teaching and learning apps and applications already exist and what they do.

I’m increasingly interested in specialized and localized AI models for education. We’re already seeing AI for specific teaching and learning purposes - lesson planning, tutoring, writing assistance, content generation, creating quzzes and assignments, … And I think we’ll increasingly see domain specific AI models eg. math, biology, history, art, … As these emerge, if the models are open, it will be possible to localize and contextualize them by language, culture, and other ways that make them more relevant and effective.

Cathy wonders about the compatibility of OE, AI and commercialization and asks, “Could it be possible for the OE community to commercialize OE for the benefit of the OE community, rather than await the threat of the corporate sector acting independently?”

@danmcguire @Jenni @JudithSebesta @cogdog @billymeinke and @Mackiwg all make some really good observations most of which I fundamentally agree with.

The way I see it AI definitely has the potential to accelerate and improve the production, discovery, use, or uptake of OER broadly. But AI is currently operating in ways that break the social contract of OE, perpetuate extraction, colonization, and imbalance of power. How unfortunate!

I keep wondering if we can articulate how AI could / should work differently. How might AI work collaboratively with the OE community in ways that align with social norms and the commons? I keep wondering if AI community rules and regulation could steer AI to be not just an industry sector and a race for economic benefit but something that also, perhaps in parallel, generates social public good of the type OE enables. I can envision rules and regulation that favour the latter over the former.

However, the current VC funded AI tech landscape is not functioning in that way and not likely to change. In addition governments seem disinterested in the social public good potential and as a result regulations are not steering AI in that direction. AI needs a model that does not give big tech disproportionate benefit.

Toward that end there is quite a bit of activity around opting out of AI. However, I believe we need an alternative that we can opt in to, something that gives us agency in an affirmative way. I’d love to see open public infrastructure for AI and associated open governance (with strong representation from OE) of that infrastructure. AI infrastructure is hugely expensive but in my mind funding for this could come from a levy placed on commercial AI (which then ensures a reciprocal benefit back to the commons). The EU AI Act advocates for nations to build their own AI infrastructure to ensure an element of sovereignty and independence from big tech. I’d love to see the OE community advocate to its governments interest in using that infrastructure for education rather than that of big tech.

AI is built on data. Its future sustainability depends on continuous ongoing ingestion of new data. Interestingly human curated data is considered higher quality than data simply scraped from the web. This is an area and practice that OE excels at. I can see a convergence of interests between AI and OE around this aspect. I definitely think OE could generate its own AI models. Such models could be higher quality and generate better results than those offered by big tech. Is there a business model around that? Maybe.

Data is a kind of currency. Two data sources continuously feeding AI currently are user interactions with AI applications and outputs generated by that AI use. One way the OE community can assert influence is by controlling how their data and outputs are used. I keep imagining a scenario where I agree to have my data and outputs used by AI specifically for social public good rather than for profit.

Over the years I’ve given a lot of thought to this fundamental question about how to generate revenue while engaging in open work. Like many of you I’ve done, and continue to do, a lot of open work that generates no direct financial return. But I’ve come to believe that it generates a lot of “value”. I wish the conversation was less around commercialization and money and more around creation of value. I believe OE offers a higher education value proposition than non-open. Direct financial benefits are but one component of that value proposition and focusing on them exclusively diminishes the other values OE generates (many of which have indirect financial benefits).

That said I don’t believe OE should ignore the financial side. When I think back to natural resource based commons such as pastures, water, and forests, those were managed as commons partly for the economic benefit of all involved. In current times those have largely disappeared. Non-profits and efforts like OE exist because of failures in the way society and the economy work. While adopting commercialization as a goal risks co-opting OE into capitalism it need not be the case. Certainly social enterprises and even B-Corp often have missions highly aligned with OE and public social good. But finding, or establishing, an organization to take on commercialization of OE for the benefit of the community requires significant startup funds and ongoing governance and operations that I currently don’t see anyone willing to provide.

This conundrum remains a topic of high interest to me. There is no magic bullet but continued conversation and dialogue like this one open the door to more possibilities. I’m all for working together on coming up with new models.

1 Like

Hi @danmcguire , to clarify what I meant about selling products for student use, what I meant were AI products (fully vetted for ethics, cybersecurity, student safety, and pedagogic value) for students to use to co-create knowledge (or other active AI-focused or AI-supported learning opportunities). Definitely not the sale of OER. I would want any products students use to be paid for by their institution (in my ideal world).

1 Like

@paulstacey and @mackiwg

Thanks for your contributions. Paul - it is terrific you are presenting on this topic in Taiwan. Please share the link once available.

Many terrific points raised and lots to consider. Not an easy topic to parse, and no clear answers.

You raise a good point about value being consider separately from revenue. Certainly value is linked to revenue but need not include revenue. Certainly nonprofits had huge value to society and are not renumerated.

The concern about “the revenue engine” - whatever that might be, not being available to those most in need because of cost is a real concern. I think of the “Red Hat” model as an example that works, and acknowledge that this particular model is unique.

More to ponder…

1 Like

@Jenni If you’re selling proprietary products you’re not an advocate for OER. I can say that my proprietary word processing program (MS Word) can be used to create OER but that doesn’t make me an advocate for OER or Open anything, especially if there’s an open source way to do the same thing.

Here’s an example of a new tool that uses AI to create OER and is open source. The OERtist Tool.


Hi @danmcguire have we met? I think it’s probably best we don’t engage in too much of a “what is open, how is it done” debate when we know very little about each other. I look forward to learning more about the AI tool you reference in your posting. Especially all the ways that the AI will be trained on expert use of the multiple first languages of Africa for primary school-aged children and how the data-driven enhancements and adaptations to student learning and teacher teaching will be structured, confirmed as authentically appropriate, and vetted by local teachers for the contexts of the learners. It sounds like a great project!

1 Like

We had a great OEG live with Dan and Kathryn Cure, see archive and lots of links

Its a very exciting project

1 Like

Somewhat marginally related, but hopefully of interest is this post in the Scholarly Kitchen

The first point of interest is the animation from a Google Research article suggesting as LLMs increase power, in this case discussing a 540-billion parameter model that the capabilities increase dramatically – certainly open to debate, but also look at rate of growth of LLMs.

But tha main part of the editorial/post is a stand counter to what most academics are asking for-- rather than trying to keep research and academic content out of LLMs Stuart Leitch suggests we should get as much content in as possible (?)

Academia has developed an amazing tree of knowledge which is arguably the most important data for Large Language Models to be trained on. The frontier foundation models are widely assumed to have been trained on a wide variety of paywalled content. There are ferocious legal efforts underway to get this content out of the training sets.

The provocation I put forward is what happens when we have models growing ferociously in capability, but we decline to train them of the very best sources of human wisdom and instead have them learn on the longer tail of less rigorously curated information, or information that is out of date. What does that do to the risk that these technologies are needlessly rough on humanity? If we succeed in getting the LLMs unhooked from the best sources of information right as they are set to have whole new sets of capabilities emerge and are being integrated into everything, how might that play out? Do we want the tech oligopoly turning to simulations to generate the training data?

If we treat this as a gifted child that may take over the world, we owe it to humanity to give it the best education possible and to ground it in the best of human wisdom. It for sure knows about Nietzsche, Machiavelli, Sun Tzu, and Clausewitz.

My provocative thought is that rather than trying to get premium scholarly information out of LLM training sets, we should fight to get it in there, on terms that are economically sustainable.

I am not necessarily agreeing, but appreciate what is described as “food for thought”-- but also, he does not elaborate on that last sentence “on terms that are economically sustainable”

Hello Cathy Casserly,

It’s wonderful to have someone with your extensive background and experience in Open Education engage with the community. Your question about the compatibility of Open Education (OE), AI, and commercialization is timely and raises important considerations.

The intersection of AI and education presents numerous opportunities and challenges. Leveraging AI in educational technology can enhance personalized learning experiences, facilitate adaptive learning platforms, and provide valuable insights into student progress. However, as you’ve rightly pointed out, the question of commercialization can be sensitive within the OE community.

Here are a few thoughts on the matter:

  1. Balancing Sustainability and Openness:

    • Finding ways to sustain the OE community is crucial for its longevity and impact. Exploring revenue streams doesn’t necessarily mean compromising on the fundamental principles of openness. It’s about striking a balance that ensures financial sustainability while preserving the core ethos of open education.
  2. Collaborative Commercialization:

    • Rather than leaving commercialization solely to external entities, the OE community could explore collaborative models. This might involve partnerships with education technology companies or developing commercial offerings in a way that benefits the community directly.
  3. Ethical Considerations:

    • Any exploration of commercialization should be underpinned by a strong commitment to ethical considerations. Ensuring that the benefits are distributed equitably and that commercial activities align with the values of the OE community is paramount.
  4. Open Source and Licensing Models:

    • Leveraging open-source models and appropriate licensing can be a way to commercialize without compromising openness. This way, revenue generation can come from support services, customization, and other value-added offerings while keeping the core educational resources open.
  5. Community Engagement:

    • Involving the community in decision-making processes regarding commercialization is essential. This can help ensure that any initiatives align with the collective values and goals of the community.
  6. Education and Advocacy:

    • Educating the community about the need for sustainable funding models and the potential benefits of controlled commercialization could foster a more receptive environment. Advocacy efforts can help shape a consensus on how to navigate this complex landscape.

Overall, finding the right balance between sustainability and openness is key. The challenge lies in devising models that allow the community to thrive financially while staying true to its open education principles. Your experience and insights will undoubtedly contribute significantly to navigating these complexities.

1 Like