Do you really need to implement AI inside your ELearning platform?

Next Story

HQ Trivia business model, how much $$$ do they get?

For the sake of the simplicity let’s omit the difference details between AI (Artificial Intelligence), Machine Learning and Deep Learning, for regular users it’s all the same buzzwords.

 

Recently me and a customer discussed development of a new eLearning system, including AI system inside . However she doesn’t know yet which exactly AI to use, but she needs it because investors expects them that AI will guide student during education process. Do they really need it? Let’s try to answer in this short review.

This request for AI is not exceptional case, nor the rare one and here is why:

This is Google’s search trend over the last years, the popularity of AI is skyrocketing, but why?

The reason is that few things came together around the same time, which are really important for that AI to function:

  1. New AI learning algorithms were invented.
  2. Computers become fast enough to handle big amounts of data required for AI system.
  3. That Big Data become more available.

 

And here is another buzzword “Big Data” – it just means a lot of data. What kind of data? For example your system contains learning process details about students who passed the courses. When they study, how much time did they spent on specific tasks, when they clicked interface elements, etc. That’s all could be used.

Ok, but does this AI popularity rise means it brought the same amount of usefulness? I hardly doubt that. Certainly it did in some specific areas, but let’s not overestimate the impact.

How it can help in eLearning industry? Basically AI systems can solve next set of problems:

  1. Classification: You have predefined groups of something and AI potentially can learn how to divide new information by those groups. Example: A. Automatically grade student by his learning process data. B. Recommend activity that should be useful in studying process as a next step.
  2. Clustering: Find similarities and group it in a set of chaotic data. Example: A. Find common mistakes that subset of students do.
  3. Regression: Kind of future prediction. You feed the incomplete data and receive quantity prediction. Example: A. Tell the student that according to his current progress he might get next results.

So how it all does work and how hard to implement it? Briefly it looks like this:

Take Big Data —-> Feed it to the AI system —–> Get the result

Is it all that simple? In some sense, it is. However there are tons of details and eventually it becomes more sophisticated.

  1. First you prepare a Big Data. Yes, you need to have a lot of data to run the AI system. And this data should be prepared in a right way. That is a job for Data Scientist. Is it time consuming? Yes, it is. How big is a big data? Roughly let’s say you need to have millions of records.
  2. Then you create your AI system and train it by using a big data. It’s not an easy task and Machine Learning experts usually do that. Again it’s time and resource consuming.
  3. You integrate the created AI system into your eLearning system workflow.

But do you really need to use the AI system in your product? As always it depends!

I think next considerations will help you make a choice:

  1. If you are a startup and you are not expert in AI then most probably you don’t need it. You can solve the problem manually or in a traditional programming way. At least until your company become profitable and grow.
  2. If you are a bigger company and you clearly know the problems you need to solve, then it’s good idea to check whether AI system can come handy. But please remember that you need to have a huge amount of data. If you not sure about your data it’s good to have a chat with Data Scientist.

For the customer I’ve described in the beginning AI system is not needed. They can mimic the AI by carefully handling study process recommendations themselves by humans. And when they polish their business idea, validate the recommendation mechanism and gather Big Data then it would be good time to check possibilities of AI. Eventually then can create a system that will work kind of Sorting Hat from “Harry Potter” 🙂

On the other hand, even if you are a startup or not ready to invest into complex AI. There are ready AI platforms available that you can integrate into your system, for example:

  1. Chatbots. Why not to run interactive questionnaire for students by a chatbot. Or it can remind students to do their homework on time.
  2. Image generators. What if AI system generate custom art images or music during educational process.
  3. Voice recognition systems. In mobile apps it’s harder to type on keyboard, why not to try voice recognition.

Do you know other practical examples of AI inside eLearning apps? I would love to hear them!

 

104 views