Generative OER, Generative Learning, and the Generative Future of Information Dissemination

Plus, Generative Archival Access?...

A few weeks ago, I had the privilege of attending a webinar by Dr. David Wiley entitled “Why Open Education Will Become Generative AI Education” (video below). The thesis of this webinar was that just as OER materials in general had altered the education norms of the 1990s and 2000s (which effect is continuing today), generative AI was becoming an equally important and so-called “disruptive” force, both as a material and an educational technology. Wiley furthered his point by suggesting that “generative OER” is going to become the next major version of OER. “Traditional OER,” or textbooks and worksheets that are openly licensed but still made for groups rather than individuals, are on their way our in the world of open pedagogy.

Several individuals in the comments of this webinar took fairly strong offense to the idea that “traditional OER” is losing favor as a result of the rise of “generative OER.” They seemed to forget that OER materials were themselves extensions of previous educational materials (copyrighted textbooks and ancillary materials existed long before openly-licensed materials). Furthermore, other educational technologies have been adopted in the past decades, and I (as a student from the 1990s to the present) have to say that almost all of these technologies had issues and vocal detractors. Many of these technologies are now standard in general education, such as computers, the internet, graphing calculators, and other tools.

Furthermore, some people accused David of acting as though people who had create “traditional” OER materials were not “as good” or “as effective” as those making “generative OER.” They forget, it seems, that for over a quarter of a century he has been involved in the open education network. He built an entire company around traditional OER materials and open pedagogy, he has made dozens if not hundreds of “traditional OERs,” and that he is one of the people who invented the concept of OER in the first place.

Now that I’ve said my piece in a battle that was not really mine to fight, let us continue to consider the ideas behind David’s presentation, and why they are so important in the broader information dissemination field, not just formal education.

Generative Tools Create Human-Centered, Personalized Interactions

Generative AI tools enable human-centered interactions, as I have discussed in multiple earlier posts. By collaborating with users rather than simply performing tasks, these tools adapt to personal preferences, prior interactions, and contextual needs. When educators integrate AI properly, it augments human capabilities, ensuring the user remains central to the learning process. For example, educators and students can use AI to build lesson plans, personalize learning outcomes, and craft creative assessments. The potential for collaboration is immense—AI now acts as a collaborator, pushing the boundaries of what users can achieve instead of merely functioning as a tool.

Generative OER

Generative OER combines the open-access principle with AI’s content-generation capabilities to facilitate new educational resource development. In Wiley’s vision, generative OER crafts textbooks, lesson plans, and interactive learning materials that respond in real-time to educational demands. For instance, a biology teacher can request a custom textbook section on genetics, and AI can generate resources filled with diagrams, readings, and self-checks. Multiple iterations are needed to make the final product, but the process can begin and even advance a few stages completely in the outputs of a generative AI tool.

In my blogpost Deliberate Creation, I talked about Dr. Law’s Rhetorical Framework of Prompting, and emphasized that users should heavily consider the audience, purpose, format, objectives, and other aspects of their desired product before they ever touch an AI tool.

Pay special attention to the section, “How Do We Prepare to Collaborate with AI in ‘Deliberate Creation?’

Generative Teaching: Educators

Generative AI now supports teachers in the classroom. One of Steve Hargadon’s major ideas is “generative teaching,” or “generative learning” when performed by students. This is not his concept, but something that initiated in the work of Merlin Wittrock in 1974. The central concept is that new information must be actively received by students so it can be combined with previous knowledge. In instructional design or cognitive terms, students must be active in order for pre-existing schemata to be altered to accept new information.

One of the best ways to do this is to have individualized learning through projects or experiences. I am a great believer in project-based learning or problem-based learning, and this goes well with both of those models.

Educators can create course-based generative AI models (whether open or not is another question, and I will write another blog post specifically about that issue) and have their students interact with them. They can prepare the AI tool with formative questions, textbook ideas to emphasize, or even have the AI tool act in place of the textbook. It could also connect ideas in the course with ideas found in external resources, such as the internet. The tool could be prompted to help students connect course ideas with things or concepts that are personally important to enhance acceptance of the new information. The key is to prime the AI tool with correct data and information and let the student and the AI tool communicate. Potentially, the students could use the tool to help them plan final course projects (or even generate first drafts or edits of those projects).

In my blog post, A (Working) Framework for AI Literacy, I argued that one of the central challenges of AI in education is ensuring that teachers become AI-literate​. Generative teaching tools automatically produce lectures, assessments, and feedback, freeing teachers to focus on meaningful interactions with students. However, educators must understand how to use these tools without relying on them to handle critical thinking. They must serve as “co-pilots,” guiding AI toward specific educational goals while maintaining ethical considerations and realistic awareness of AI’s capabilities and weaknesses.

Generative Learning: Students

That’s all very well and good for educators, but what about students? How can they use AI tools to further their learning, or must they only use what their educators give them?

In his webpage for a higher ed course on AI and instructional design, David Wiley includes prompts for AI-moderated quizzes on course topics. This is, admittedly, instructor directed, but the direction is relatively loose. The tutoring session can include any sort of communication so that the student understands the concepts before they answer the quiz questions.

Students can take the initiative and put course material into an AI tool to create their own study tool without instructor direction (although permission should probably be granted). For security and data purposes, I would recommend that students use open-source models and a platform like LMStudio or GPT4All, which I will discuss in another post. However, I have created multiple Custom GPTs for this purpose as well.

There is an option to prevent general model training on Custom GPT interactions when you upload PDFs. While the “training” that AI tools undergo does not keep any of the data or information contained in documents or datasets, you do not want to provide personal data (or ways that a tools can extrapolate that data). Speak about your hobbies, interests, or personal preferences generally instead of giving details.

Generative Information Dissemination

This is essentially using the “generative learning” process of the last section for topics for which we have no formal learning obligation. In other words, “generative AI for professional development,” or “lifelong learning,” or whatever you want to call it.

Generative information dissemination can take many forms. This could be directed or encouraged by an administrator (you could put procedures or HR policies into a bot and have employees talk with the AI tool) or employee-initiated (you can roleplay a conversation with your boss or simulate a difficult customer-service interaction or negotiation with another stakeholder). You could use generative AI to give you quick quizzes on info or data you need for an upcoming meeting, or review your slideshow script for your monthly report. You can also just use AI as a conversation partner, which is what most people do.

Generative AI has altered how we create, share, and consume content. It can generate reports, summaries, and even some sections of research articles. While we need to exercise information literacy (I heavily encourage you to use the SIFT Method) with AI tools as much as with regular websites, if we have all of the aspects of AI literacy we can navigate information with the aid of generative AI tools. I have created CustomGPTs for this purpose, and many of them can be found at libraryrobot.org. Check out the Search Query Optimizer, especially.

Steve Hargadon likes Perplexity for looking at information and analyzing data, because Perplexity automatically had access to the internet built in to its processes. Furthermore, Perplexity always gave links to its sources. ChatGPT 4o and o1 usually give links to their sources, and I think they are more flexible and conversational.

As I wrote in a post a few months ago, We Haven’t Gone As Far As We Can Go,

“Even if you are writing a textbook, please write it in your own words. Use genAI to help you come up with the structure, the wording of the learning objectives, even discussion questions (that, again, you edit). But please do not rely on it alone for information about any topic or skill. You will only be putting yourself at a disadvantage, especially if you are required to work with others who have learned the skill or topic with traditional means and structures. I am not talking about bias or prejudice against personalized learning (that is a conversation for another day). I am talking about the validity and usability of the information given by GPT 4o when it is not connected to the internet or given structured and heavily-revised prompts.”

If you are asking an AI to disseminate information to you, at least provide it with access to authoritative sources, whether that is through PDFs of reliable sources, prompts to specific websites or specific types of sites, or give it a dataset that contains reliable information and data. Once you have done this, and still used the SIFT method, have another person analyze the conversation. This is similar to having an editor read over a chapter or article, which everyone should be doing in the first place.

Generative Archival Access

It seems like a large number of people in the general library field have discovered and are beginning to harness the capabilities of generative AI tools. At the GAIL conference, there were no fewer than three AI library chatbots being discussed. There were even more library chatbots discussed at the AECT 2024 International Convention. AI library chatbots are alive and well, if they are still in their natal stage.

Not much, to my knowledge, has been done to bring generative AI’s information communication abilities to archives researchers.

Generative AI’s potential extends to archives, a traditionally human-centered space. In Archivism in the Age of AI, I explored how AI can enhance archives by automating metadata tagging, organizing collections, and interpreting handwritten documents​.

Pay special attention to the “Patron Services” section of the article above. This section is an expansion of that part, especially the concepts of selecting and describing entities and expanding the discovery abilities of patrons.

Generative AI could take this a step further by creating collaborative archives where users engage in real-time dialogue with the system. Imagine an archive research tool that not only retrieves documents but also generates interpretative summaries or curates content based on the user’s research needs. An archive tool that connects archival materials in local collections to others from around the world and to non-archival materials such as websites or books. This could be as simple as including a prompt in a conversation to search another website catalog for materials with related entities:

“This letter from James Fenimore Cooper talks about wilderness scouting, what other archival materials in the catalog are related to this topic? Could you also search the UC Davis special collections for materials related to “wilderness scouting? Now, let’s search .edu or .org websites for information about the main parts of wilderness scouting.”

AI enhances accessibility and makes archives more dynamic, but human expertise remains essential in guiding these processes. As with researchers in a physical archives, the human must have previous knowledge about either the topic they are researching or the entity or collection they are wishing to peruse.

As We Engage with AI, We Must Cultivate AI Literacy

Whether it involves generative OER, teaching, or archival access, AI literacy remains crucial. Users must understand how AI operates, its ethical implications, and its limitations. In A (Working Framework for AI Literacy), which I linked above, I outlined the competencies necessary for effective AI use​. We must train a generation of learners and professionals who can scrutinize AI outputs, using the tools wisely while ensuring human oversight. Without AI literacy, we risk atrophying our ability to use them responsibly.

References

Law, J. B. (2024, April 1). Rhetorical “Rhet shots:” mastering the recursive dynamics of the rhetorical “Rhet” shot prompt engineering method in collaboration with your AI assistant. Prompt4all. https://prompt4all.wixsite.com/prompt4all/post/rhetorical-rhet-shots-mastering-the-recursive-dynamics-of-the-rhetorical-rhet-shot-prompt-engin

Van Campen, K. (n.d.). Library guides: Evaluating resources and misinformation: The SIFT method. The SIFT Method - Evaluating Resources and Misinformation - Library Guides at UChicago. https://guides.lib.uchicago.edu/c.php?g=1241077&p=9082322

Wiley, D. (2024a, January). Generative AI for instructional designers. Generative AI for Instructional Designers. https://ai4id.org/s24/llms/#:~:text=Weekly%20Activity%3A%20Part%202

Wiley, D. (2024, September 24). Why Open Education Will Become Generative AI Education. YouTube. Link above.