AI in e-learning isn’t theoretical—it’s already here.
It’s hard to write about AI in e-learning without feeling as though you’re engaging in the sort of speculation usually left to science fiction authors, especially as we’re used to thinking of it as something potentially sinister or frightening. However, whether or not we’re aware of it, AI has become a regular part of our day-to-day lives.
Part of the disconnect, for many of us, is that the AI that exists simply doesn’t look like self-aware, human-passing androids that writers like Isaac Asimov, Philip K. Dick, or Arthur C. Clarke led us to expect. Instead, what we have are sophisticated software programs that are very good at finding patterns, and then adjusting behavior to those patterns to deliver a fairly limited range of results.
That doesn’t sound as exciting (or as scary) as your typical sci-fi novel, but it does open the door to some very interesting use cases that are being applied today in the world of online education.
We’ve written before about how these technologies might be applied. Today, let’s look at a few companies who are already using them.
Duolingo is probably the most prominent example of AI being used in education today. They’ve been very public about the research behind their language learning app, to the extent that the advances they’ve made in applying machine learning to language learning are a part of their branding. They even have a portion of their site dedicated to their research.
How are they applying this research to their courses? First of all, Duolingo’s AI personalizes courses, adapting itself to each learner’s strengths, weaknesses, and preferences. It will pay attention to what vocabulary the learners knows, which grammar examples they struggle with, and what content they seem to enjoy.
Duolingo’s AI is also uses natural language processing to create chatbot experiences that give learners an opportunity to practice conversation in real-time. It gives language learners the opportunity to practice their skills and gain confidence before they have to step in front of a real person.
Like Duolingo, Thinkster uses AI to deliver personalized math tutoring to K–8 students. Learners begin by taking an assessment test, and then the AI can customize the questions based on the learner’s knowledge level, and how they engage with the material.
What is interesting about Thinkster’s approach is that it combines AI with coaching from real math tutors. This means the personalization isn’t just happening for the students—it’s also helping to prepare tutors to give more targeted lesson feedback. The result is that teachers spend more time focusing their attention on material learners actually need.
While learning personalization is a great use for AI, Querium takes a different path. This virtual tutoring program analyzes the steps learners take in solving a STEM problem and provides immediate feedback about what that learners is doing right or wrong. This prevents the learners from absorbing a wrong way of answering a problem while sparing teachers from an overwhelming amount of coursework to correct.
What’s so special about the use of AI here is that, in order to provide the right kind of feedback, it has to understand input data from the learner that might not take the same form every time. This is much more complicated than simply taking a structured response from a set list and providing feedback, but it also allows for more accurate instruction.
4. Alta by Knewton.
Alta, a new product from the higher education brand Knewton, uses adaptive learning to identify gaps in learner knowledge and then fill them in using high-quality learning materials selected from its own databases.
In this example, the software is a study guide, identifying and then filling knowledge gaps. Applied differently, it can also help companies maintain training so that employees can stay up to date with emerging skills or regulatory requirements.
To recap: we’re seeing 4 types of AI currently in use.
While there are other examples out there of online educators using AI, they almost all fall into the use cases listed below.
- Natural language processing. Used in language learning, but also with major accessibility applications. Tends to be faulty around children or people who are multilingual.
- E-learning personalization. Adjusting course material based on learner use and preferences.
- Virtual tutoring. Grading assistance to identify and correct learners errors.
- Adaptive learning. Proactively identifying and resolving gaps in learner knowledge.
Many of these can be combined and used together to form a richer AI-powered learning environment. And while AIs frequently need training to get to a point where they can act with intelligence, the more a program can link these various types, the more they can accomplish with them.
As AI joins the software-as-a-service (SaaS) economic ecosystem, it will become more accessible for educators.
While many examples of AI in e-learning are still concentrated on large businesses with lots of budget to invest in these technologies, the emergence of an SaaS market for AI (AIaaS) is rapidly opening the field to further experimentation.
The growth of AI in online education is excellent news for everyone. The more educators can experiment with the technology, the more new and innovative ways they will find to put it to use. And as these use c
Not long ago, many technologies that underpin the Internet required specialized skills to use. We’re not far from a future where AI is just as common a component of course creation as video.