So, I was recently asked, “What exactly is AI and what does it mean to say AI in education?”

Hmm, good question. I’d never really thought about it. Sort of like asking someone “What exactly is the color blue?” How would you describe it to someone who had never seen it? It, blue that is, is one of those things that you just know what it is when you see it regardless of the shade of blue you’re looking at. Right? I mean, you know Carolina Blue (Go Heels) when you see it. And you know Duke Blue (boo Devils) when you see it too. But it’s hard to describe it without referring to something else. For example, Carolina Blue is a light blue like the sky on a sunny day. Duke Blue is a darker blue, closer to Navy (the color – not the ships of the line).

Artificial Intelligence (AI) is both like and not like blue. You may think you know what it is when you attempt to describe it, but you may not recognize it when you see it (e.g., “because you did x, you may like y” … that’s AI).

In the educational space, AI falls squarely in the domain of educational technology (EdTech). EdTech is a type of tool that the teacher/instructor/trainer has at their disposal to help facilitate the transfer of knowledge (i.e., learning). Because it is a tool it must be used in conjunction with a human (the teacher). This human-in-the-loop (HITL) approach with the interactive tool of AI (Wang, 2019) should provide a power couple that can tackle the plethora of issues that come up in any classroom while doing so with up-to-date and just-in-time information to handle whatever comes up.

“Can something be described as AI without machine learning done by the technology independent of a human?”

Clark described AI as an idiot savant (Clark, 2020). AI-enabled cyber or cyber-physical systems are amazing things that can operate on their own with just minimal instructions. But there is the catch – “with … instructions.” Something must put the AI to work. Many automatic systems exist that do not have official “AI” in them (e.g., your flush toilet automatically refills the tank and then shuts off the water, but I’d dare you to find anyone sane who claims it is an example of AI at work). AI does not need to have explicit instructions telling what to do or when, but it does need someone to look after it because it has no idea of when or where it has gone wrong or gone too far and could be wreaking havoc. Bless its heart, it’s just not that smart. Clark describes the robot vacuum that engages, rather than avoids, the pet waste and drags it all over the house because it does not know it should not be doing that. So, while it can run independently of a human for some things and for some time, it cannot be left alone indefinitely.

What qualifies and what does not?

For AI to be used in education, it must contribute to the goal of educating the students it is intended to support. If the AI is not used specifically for promoting a means to an end of delivering on the learning outcomes, then the AI is not “in education” and is not “learning technology” at all, it is just collocated in the classroom. Just because it is used in the classroom does not mean it is “learning technology.” Take for example the room lighting. It is technology that is used by the teacher in the classroom to help facilitate the learning exercises because they are much easier to do when you are not in the dark, but it is not “learning technology.”

References

Clark, D. (2020). Artificial Intelligence for Learning. Kogan Page.

Wang, G. (2019, October 20). Humans in the Loop: The Design of Interactive AI Systems. Retrieved from Human-Centered Artificial Intelligence – Stanford University: https://hai.stanford.edu/news/humans-loop-design-interactive-ai-systems