What We Lose and Gain with AI in Education


Space Odysseys  

We all know what is gained by adding AI to the learning process. But what do we lose in the same process?

In 2001: A Space Odyssey, the conflict between HAL (the onboard AI) and the astronauts is based on a basic malfunction/misunderstanding of a directive, and that is the most frightening part. HAL is not focused on the needs of the people in its care but rather the mission it was given. David, the hero of the story, eventually gets around all that, but it is rough going; AI is smart. 

Yes, all of this is science fiction – entertaining but ultimately not based in real world experiences. However, the ideas are especially relevant as we see the flood of AI into consumer products and everyday experiences — including in education.

The lack of human understanding and care that an automated processor may have and how that could create specific consequences for the learner are real and worthy of consideration. It becomes important when we start talking about ability of machine learning to automate, predict, classify, generate, and understand based on learner behavior and performance data.

The Future Is Now

This concept of automating processes in education isn’t just theoretical. It’s happening right now. Many in edtech are working on this (even I did for a long time), creating technologies that use machine learning to do a wide variety of jobs usually performed by teachers. The argument goes like this (and it’s not necessarily wrong):

If certain mundane, form-and-file parts of the education system could be automated, that would free up a lot of time for teachers to do other, student-centered activities.

Yet, that may not be where automation in education ends. Even now it’s creeping into other areas like lesson prep, grading, and mentoring. While there are many, many benefits to automaton in edtech, there are also drawbacks we want to keep in mind, especially as it relates to the role of teachers in the learning process. 

3 Areas Where Teachers, Not Machines, Are Needed

1. Student Engagement

Kids might think AI is cool, but the novelty of it will wear off. Then what happens? Boredom, distraction, and inattention set in. Machine learning is not currently at a place where it can tell, just from indirect human cues (like the combo of facial expression and tone or gesture or fidgeting or specific word choice) when a child is no longer engaged. A teacher can. They can pick up on these things, in face-to-face or virtual instruction, and respond in real time. Student engagement is important for learning, and teachers are still the best tools for judging engagement. 

2. Personalized Learning Plans/Actions

Edtech automation is now trying to extend into personalized lesson planning a bit. Sure, at first glance individual lesson plans may appear to be scalable and quantifiable, but it’s not always one-size-fits-all. Teachers come to personally know their students. That personal knowledge and interaction allows for individualized planning that cannot be mimicked by technology just yet. Teachers can see/hear/read how a student is progressing and adjust plans according to various assessments that involve quantifiable data AND basic human interaction with that student. Both are needed to create solid lesson plans for the whole class and the individual students. 

3. Emotional Intelligence and Social Skills

Education is not just about academic knowledge. It’s also about learning to socialize, learning how to interact with people, and learning to get along with, work with, and understand others. It’s about emotional intelligence and social skills. As smart as AI is at this point, it doesn’t have that. These skills still require good teachers to be present for students in a thousand tiny ways that involve interacting with an older human being who has a bit more experience out in the world. AI can’t model human behavior fully, but teachers can. 

All three points really hit on the same idea, just from a different angle: you need people in education to have important human interactions. That’s something to remember as more and more tech is placed in classrooms. There are many good things automation can bring to edtech and education. I’m not saying we shouldn’t be working on automation in edtech. If it helps teachers and students have more time together, or teachers have more time to focus on student needs, that’s great. However, it can’t do everything and we shouldn’t try to make it fit every role. Teachers are still very necessary to a good, well-rounded education because they will focus on the real, human needs of the student, not just the overall plan or directive. Let’s not go trying to replace teachers with machines. At least not until they develop a real life Data à la Star Trek: Next Generation.


Leave a Reply

Your email address will not be published.