We’ve shared ideas for how to make AI largely irrelevant within a college classroom, but also want to share our thoughts on making use of Artificial Intelligence as a learning tool. There are hundreds of different types of AI, with more being created daily.
For the purposes of this essay, we’re focusing most specifically on generative AI models which create media (text, images, video, etc.) in response to prompts. The most readily available for text (right now) are ChatGPT, Bing Chat, and Bard, all chatbots. These can be used to generate entire news stories, essays, and even legal briefs. These bots are trained on data sets and, when prompted, create on the basis of that data. A bot fed Chaucer, for instance, would be able to generate a new Canterbury Tale that followed the style and linguistic choices of Chaucer’s work. Probably. The larger the data set, the more the bot has to work with and the more ‘original’ the generated content might seem. And the key word there is “seem”, as AI-generated text is as likely to demonstrate perfect grammar as factual errors.
The changes in AI are coming hard and fast and how we talk about and control/limit/use AI is in a state of constant flux. The current situation reminds us very much of a quote by the political philosopher Antonio Gramsci:
“The old world is dying, and the new world struggles to be born: now is the time of monsters.”
Anything substantive we say about the legality or utility of AI will probably no longer be true in six months (or six weeks or even six days). Nevertheless, there are some ways to think about AI that can help us be wise in dealing with it as educators.
To deal with it, we must admit it is here.
Even if it is a monster under the collective bed, we are not better off or safer keeping our head under the covers. If you’re someone who chooses not to make use of AI personally, you’re probably using services that depend on AI. If you have an AI-free classroom, it’s still an even bet that your student use AI, enjoy using doing so, and are becoming increasingly adept as users. And not necessarily “in the good way”.
There are arguments that AI is, in many respects, simply another iteration of information access, building on the tradition of written language, the printing press, and the Internet. We are old enough to recall the panic of many when the World Wide Web became available to students. But we also recall the joy we felt at being able to offer the collected works of Shakespeare to our students FOR FREE. And of course, the ‘Net eventually became a virtual bazaar where students could buy exam answers and research papers.
We suspect something similar probably happened when written language was created… A group of youngsters probably taught themselves to read (which meant they didn’t have to remember stuff). No doubt this upset the adults who were trying to teach them to listen and remember things. And with the printing press, pages of the first books were probably surreptitiously removed by young people, giving them access to information their teachers though they shouldn’t have. (When there’s one book and the teacher has it, it is the teacher’s edition by default.)
Is AI just the next iteration? Maybe. But even if it isn’t, we need to consider the action-reaction tendency that education tends to have when it comes to new technology. In each instance, we have found ways to co-opt the technology as a learning tool, even as we try to limit our students using it to avoid having to learn.
While governments and universities struggle to write policies for AI use, educators need to figure out strategies for accommodating this “reality of monsters” while still striving to maintain the best possible learning environment where students can learn as effectively as possible. We believe that this is possible if AI is treated as a LEARNING TOOL.
We define a learning tool as an instrument designed to be used by learners to provide a structure for growing learning skills and behaviors and/or systematically collecting and thinking about key information.
While (most) AI isn’t specifically designed to be used by learners, we believe that educators can design course and activity structures that specifically constrain and harness AI so that students use it in ways that target their learning skills and help them collect and work with information.
The Faculty Guidebook offers some useful Principles of Effective Learning Tools. If we’re using AI as a learning tool, we need to ensure that its use abides by these principles:
- Effective learning tools must be learner-centered.
- Learning tools should increase learning outcomes and productivity.
- Learning tools must be appropriate for the level of the learner.
- A learning tool should be engaging.
- Learning tools should enhance the development of learning skills.
- The learning curve for using a learning tool should be appropriate for the learning benefits derived.
- Learning tools should be adapted to accommodate diverse learners and situations.
- Consider that learning tools can provide a framework for metacognitive thinking and learning.
- Learning tools must align with and support assessment and grading systems.
- Learning tools should be designed for independent use by students and transfer to new applications.
We found ways to use written language, books, and the internet as high-quality learning tools; we suspect the same is possible with AI. In other words, if the monsters are here, let’s use them to help our students learn. (At the very least, we need to teach them how to successfully navigate a world full of monsters.)
