All over the globe, learners – whether school going children, college students, professionals or those seeking to acquire new skills – are looking for faster, better and more engaging ways to learn. A USD 100 billion dollar industry has sprung up online around these needs, providing a multitude of ways learners can get the knowledge and skills they seek using modern technologies and the internet.
Just as the marketplace and investors are getting comfortable with business models that build products, platforms and services around online learning, Artificial Intelligence (AI) has quietly been redefining the entire market over the last few years. That is what makes data-driven approaches to learning really hard.
The application of AI in learning can broadly be classified into automation of (parts of) the learning experience itself as well as gaining insights from the process that can be ploughed back into it for improvement.
Here are 5 advancements in this space that insiders are watching:
- Chatbots: Digital learning today gives learners content when they need it, where they need it. The content could be video, audio, textual or graphical. Learning could be integrated into quizzes, games or case studies. However, one critical element is missing: what if the learner has a question? Chatbots provide an important method to inject the element of interactivity in learning, providing instant and meaningful feedback. AI-driven chatbots go a step further than conventional ones: they have personalities, they can learn from databases, past interactions and online resources, they can empathize with learners – in short, they can be experts! AI-powered bots use techniques such as semantic analysis to gauge whether the learner’s sentiment is positive, negative or neutral; and tailor appropriate responses to encourage learning. Think of chatbots that interact and ‘prod’ learners – to finish homework, to work on their learning paths or tracking their progress. A net nanny that motivates the learner to keep at it!
The proof is in the pudding: Chatbot market size has grown from USD 700 million to over 3 billion in the last 5 years!
- Recommendation engines: If you’ve taken suggestions from Amazon or Netflix to decide what movie to watch, you’ve interacted with a recommendation engine. In learning, such engines look for patterns of behaviour based on learners’ actions and persona (e.g., job role) and together with insights on behaviours of similar groups of learners, makes recommendations on what to learn and how to learn. Many industry experts call recommendation engines a natural fit for online learning. Although very nascent, work on such engines is ongoing. Such engines can be a powerful lever to personalising learning experiences.
- Personalisation: Adaptive learning is the natural outcome of using AI-based recommendation engines in learning. Personalisation entails modulating the actual learning experience based on recommendations. The challenge here is to produce the content required for all the permutations and combinations personalisation throws up. Most e-learning caters to the ‘mass’ – the vast majority of people that lie in the middle of the bell curve – leaving other types of learners out in the cold. Imagine catering to different levels of learners – from beginners to experts – for the same topic. Personalised content throws up a huge opportunity for content creators.
Personalisation can have a huge effect on learner motivation. For instance, learners don’t have to waste time on irrelevant topics and can zero in on what they need, when they need it most. Self-paced learning is also a big benefit that personalisation affords.
- Insights: Current L&D designers use metrics for learning such as time spent on a piece of content, course or activity, drop off rates on videos, assessment of quizzes and so on. However, none of these are accurate gauges of learning outcomes. What L&D bosses (and parents as well as teachers of younger learners) want to know is this: are the learners exhibiting a change their performance and behaviours due to the learning? This is where AI/ML can help. ML-driven learning systems can use insights derived during learning to address hidden learning gaps, make post-learning interventions and refreshers to reinforce learning. Learner actions can be tracked to report an actual change in behaviour.
- Predictions: All the data and insights gleaned during the learning process can be deployed in the most powerful way to predict learning outcomes, both positively and negatively. This allows learning designers to design interventions for different types of learners, making learning more effective. Predictive learning can mean better resource allocation (think optimal training time and training payroll hours), generation of unique learning maps for learners (think automated content scheduling and delivery), improved ROI (think to deploy the right human trainers at the right times) and a host of other benefits.
If you’re an innovator, industry watcher, L&D leader or an entrepreneur in the space, start developing a game plan on how best to integrate AI/ML into your ecosystem. Be realistic about the advantages and limitations of the technologies. Much more is yet to come: we’re in for some very exciting times for learners and for the industry.
Guest Author
The author is the CEO and Co-Founder of Get Me A Course (GMAC) which is an edutech company that provides tailored solutions for learning, mentorship, and employment.