The 6 hottest learning trends for 2018 – Part 2: Artificial Intelligence (AI)

"Emergence of mysterious Black Box" by Thierry Hermann is licensed under CC BY 2.0.

This is Part 2 in our series on the 6 hottest learning trends for 2018. In last week's blog post we talked about augmented reality (AR) and virtual reality (VR). Today we continue our dive into the 6 trends with a look at Artificial Intelligence (AI). If you like the ideas and strategies we’re sharing, be sure to post a comment below, and stay tuned as we continue this series.

What is artificial intelligence?

Depends who you talk to, and what they’re using it for. Artificial intelligence isn’t one particular thing—it’s a category of technologies designed to perform tasks by simulating human intelligence. You’re probably using many such tools already—speech recognition apps such as Siri, call-centre chatbots and translators such as Google Translate all have AI as their backbone.

These tools are “smart” in the sense that they can perform tasks requiring human brainpower, often more quickly and comprehensively, although with a narrower focus. For example, AlphaGo famously beat 18-time world champion Go player Lee Sodol in four out of five rounds of Go. But AlphaGo does nothing but play Go; if you sent it to the World Scrabble Championship it would come home empty because it has a narrow focus —  hence it’s considered “narrow AI.”

A subcategory of AI is machine learning (ML), which works on the principle of giving machines data and then letting them learn for themselves. Examples include facial recognition software that “learns” to differentiate faces when it’s fed thousands of different feature sets. Or predictive software that monitors your shopping behaviour at the grocery store, then pushes coupons to your phone based on whether you bought Corn Flakes or Cocoa Puffs. Again, this type of machine learning is considered “narrow AI” because it has a narrow focus.

The opposite of narrow AI is “general AI.” Many experts contend that general AI is still science fiction. Even self-driving cars and other sophisticated AI systems don’t exhibit anything like general human intelligence; they are limited to their specific, narrow functions. As cool as it is that Netflix offers you exactly what you want, Netflix can’t go beyond its function—it’s not going to make you popcorn to go with that movie.

So how does AI fit into learning?

The best thing about AI is its ability to take over the most time-consuming chores associated with producing effective courses. AI can do the heavy lifting when it comes to collecting data, finding correlations and drawing inferences, building learner profiles, evaluating content, generating test questions in a variety of domains, and adjusting course content to fit individuals’ learning needs. By leveraging AI, course designers can apply their energy to what they do best (and what still eludes AI): developing awesome content.

What kinds of AI functions can be used in learning?

Chatbots can assist learners by directing them to particular resources they haven’t accessed yet. They can identify gaps in learning and prompt the learner to visit particular courses or sections of a course. And they can reinforce skills by prompting the learner to practise them or apply them in different scenarios.

Training can be personalized to the learner. Adaptive AI can make course changes based on the learner’s performance, strengths and weaknesses, aptitudes and even job performance or productivity data. A course needn’t be static—if a learner has no need for certain sections, AI can adjust the curriculum so they get skipped, or add other sections that apply specifically to the learner’s job description. And based on learner preferences, tasks can be tailored to be more visual, auditory or kinesthetic.

Data can be tracked as part of a company’s overall talent development and retention strategy and used to inform onboarding, everyday performance and incentives.

Large quantities of data can be amassed and used in collaboration across an industry. For example, medical imaging data is used to train physicians to spot various conditions; a robust dataset enables better diagnosis and collaboration among physicians when second and third opinions are needed.

Engagement may be boosted because content is tailored and paced for the individual learner. Personalized goals can be created, and chatbots are getting better all the time at providing encouragement.

More personalization means less “seat time” and training payroll hours, which means greater ROI.

Peer-to-peer learning can be facilitated by AI; learners and mentors can be matched based on their capabilities, as captured in the database. This can increase motivation and course effectiveness.

The curriculum can be self-correcting. Metrics show you what’s working, what’s not, and the platforms learners are using to access the materials. This helps you know how to tweak your content as well as how it’s delivered.

As far as automatic content creation goes, it’s still a pipe dream. But in certain domains, such as mathematics, AI can generate almost limitless test questions.

 Who’s using AI?

AI is everywhere! From car manufacturers using robots to assemble cars, to online retailers making purchase recommendations, to robo-investors, to you with your very own smartphone. All around us, things are getting smarter.

 Using AI in your learning

• Plenty of elearning software already uses algorithms and automation features; they’re built right in. Systems that are limited long in the tooth may benefit from third-party add-ons to optimize AI capabilities.

• If you’re getting started and haven’t settled on a system yet, be sure to research Learning Management System (LMS) platforms to see what they can do and how they can be adapted or optimized.

• If you’re already producing courses, be sure to look at the data you’re collecting from your LMS. See if you can enhance it with additional data—from your company website or social media pages, for example, or through survey results and job-performance data that you’ve collected over the years.

• Understand that AI is ultimately not that intelligent yet—and the jury’s out as to how smart it can get. For the foreseeable future, humans will be needed to perform complex tasks, especially if they’re social. AI probably won’t help you with those “human” functions, but it can free you up by handling information-heavy tasks such as metrics, data collection and analytics.

• Have a game plan. Think about the AI functions that are useful now, and how they can evolve along with your course design.

Is AI here to stay?

You bet it is. Increasingly, organizations are seeing the financial benefit of AI. Studies show 40% of employees who receive sub-par job training leave their positions within one year. That makes it critical for companies to provide engaging, meaningful training to help retain their people.

According to Deloitte, 84% of global executives rank employee learning as important or very important. According to a study conducted by IBM, every dollar invested in online learning results in a $30 increase in productivity. The Research Institute of America found that eLearning increases retention rates by 60%. AI can play a part in making training “sticky” through better engagement and more insightful metrics that allow companies to course-correct quickly and get more productivity from their employees.

What’s your thinking about AI? Is it the stuff of sci-fi, or is it already a part of your toolkit? Leave us a comment below (or contact us directly) and tell us what you’re doing!

Click here for Part 3 in our blog post series on the 6 hottest learning trends for 2018.