Thursday, April 25, 2019

Is AI A Game-Changer For Education (International Edition)?

Sometimes it's informative to see how some of this stuff is playing out in other settings. A post on Entrepreneur India makes the claim that "Artificial Intelligence Can be a Game-Changer for Education, Here are 5 Reasons Why" and its five arguments are, well, intriguing.

The post is from Vishal Meena from the start-up MadGuy Labs, an Indian on-line test prep company that promises to prep you for the tests for all sorts of government jobs. Meena has a degree in chemical engineering, but he likes doing the start-up thing, and has previously launched companies involved with bikes and with tourism. So he comes directly from the modern tradition of "You don't need any education background to be an education entrepreneur."

Meena also seems to lack a certain level of fluency in English. I bring this up not to make fun or to get all ugly American with people who aren't from around here, but because Artificial Intelligence applications for education have to be fluent in the language. They have to be. If you're going to sell me the eleventy billionth hunk of software that can supposedly assess writing, then it had better be plenty fluent in the language and capable of telling the difference between effective writing and tortured gobbledeegook. If you are an AI entrepreneur who wants to sell to English speaking people, then your AI had better be fluent-- and if you can't even use it to check your own work, then I'm not interested.

So-- the five reasons that AI is going to change the education game. Let's see how many of these seem familiar

1) AI to help in personalized learning.

In a classroom comprising a high number of students, attending every doubt of each student remains no longer feasible. Artificial Intelligence can help in developing personalized learning which can mitigate individual doubts and thus enhance their performances.

This was worth the price of admission for me, because here's a connection that totally makes sense, but which Personalized [sic] Learning fans hardly ever admit. Adding algorithm driven mass customization software to a traditional classroom doesn't make a ton of sense, because your software really can't do a much better job of personalization than the human teacher. But if you're talking about a classroom chock filled to the rafters with students-- well, now the human teacher has no hope of making a personal connection with individual students, and suddenly the algorithmic software looks like a relative improvement.

Also, I love that throughout this list, Meena will talk about handling student "doubts" in the classroom, which is not the way we say it in the US, but it's kind of right, isn't it? Teachers are in the business of removing students' doubts about the material and their own understanding of it. I really like the idea of students who raise their hands not to say "I'm confused" or "I can't understand this," but instead, "I have doubts."

2) Adaptive test prep

The software gets you ready for the big test by ramping up the challenge. It's refreshing to encounter someone who doesn't shy away from admitting that what he's working on is computerized test prep.

3) Addressing vernacular need

For students in vernacular learning, real-time translation ensures that the medium is democratized thereby also incorporating maximum students to enjoy the fruits of such technological advancement. It will also significantly reduce the cost of content production.

That's the whole explanation. Vernacular? I don't think so.

4) Automatic doubt solving

Clearing of doubts is fundamental to the process of learning. Artificial Intelligence engine can successfully read the problem statement and suggest a possible solution to the learners.

Again with the doubts. I'm curious how far the solution suggesting would go with an AI. In US AI plugging, we don't hear a lot about the computer providing hints or help.

5) Interactive gamification

Artificial Intelligence is instrumental in devising different tools and techniques which are immensely effective in teaching highly complex concepts in a simple and lucid way. This includes incorporation of sophisticated and useful illustrations, virtual reality and artificial reality tools, which can streamline intelligent concepts into accessible cognitive models.

Is AI good at coming up with different tools and techniques for teaching? Because I would have assumed that it can only use the tools and techniques that were programmed into it, and the programmers can only program the tools and techniques that they know about. This is one of the problems with AI-fueled personalized [sic] learning-- you're putting computer technicians in charge of educational decisions.

6) Automation of grading activities

No, I didn't mess up-- this article about the five ways AI is an education game changer actually lists six ways.

Grading work is tedious (particularly if your classroom is gill-stuffed right up to the rafters), but technology can check the answers for you, at least on items like multiple choice and fill in the blank. Meena assures us that before long, AI will start grading long answers, which is of course the same prediction that technologists have been making for decades now. Still hasn't happened, but any day now.

Bottom line? This Indian version runs right at what US advocates dance around-- the main benefit of algorithm-controlled mass customized computer-delivered education is that it will make it possible to put one teacher in a room with a few hundred students.

Meanwhile, MadGuy Labs has apparently only raised about $150K, so if you still want to get in on this ground floor, I think the opportunity is there.

1 comment:

  1. Without Benadryl in the water fountains this idea has no chance with real adolescents.