As D. Wiley explores the evolution of open educational resources (OERs) in this video, he argues that the next inevitable step in the right direction will be the eventual replacement of OERs by generative AI, which can help with a wide range of tasks.
Wiley seems to think that perceived hurdles regarding the quality and sustainability of generative AI can be overcome in a similar way to how those issues were overcome in the use of OERs. Ultimately, Wiley proposes to embrace generative AI as a worthy descendant of OERs because it has the potential to help reach universal educational opportunity.
While I was watching his presentation, I was thinking of my fellow teachers who uniformly use ChatGPT as a panacea. In my opinion, this approach is irresponsible because it trivializes learning material and reduces its educational value. I believe that for generative AI to be truly education-promoting, it must be of high quality (well-trained), and the teacher must be able to understand the power of a well-written prompt, which is what Wiley mentioned in his lecture, as well as the need for evaluation and correction by a teacher.
For example, rather than tasking AI with creating test practice reading comprehension questions, it might be more sensible to choose a text and prompt a more reliable AI, or perhaps college endorsed, to create reading comprehension questions based on that particular text. On the one hand, this gives the teacher greater control over the scope of grammar, vocabulary and overall reading difficulty, and on the other hand, it ensures greater AI accountability because of the actual text at hand, which gives the teacher greater power to control and fine-tune the final product.
Image Credit: mwitt from Pixabay

I don't know if AI will replace OER, but the research I've been doing to create some materials to train on basic AI literacy, the specifics of prompting is a very crucial aspect of using AI. As we use AI, I know I've gotten much better at prompting and refining prompts as I go. That is a skill. I appreciated your example of using a sample text to create reading comprehension questions vs. just asking for reading comprehension questions. Those two approaches would create VERY different responses. As they say, the devil is in the details. It is fun to ask AI what a good prompt should entail. It can't ready our minds. At least not yet. But I do love how it tries to with the follow up questions. Anyway, I appreciated the example.
ReplyDelete