The recent rapid growth of generative AI in business has created an unprecedented demand for continuous learning and upskilling. Learning and Development (L&D) professionals face the challenge of delivering learning solutions that can keep up with the pace of change. The traditional one-size-fits-all program for everyone can't address each learner's unique needs, goals, or preferences.
AI-driven adaptive learning is the solution that can bridge this gap, using the power of artificial intelligence to create customized and dynamic learning experiences that adapt to each person in real time.
Systems built on generative AI and machine learning can analyze large amounts of data—learner profiles, preferences, behaviors, and performance to generate the ideal learning path for each person. It's now embedded in the most advanced learning platforms; these systems provide continuous feedback to support learners and instructors, paving the way for better outcomes and higher satisfaction.
Education has been far ahead of business. Colorado Technical University began piloting adaptive learning in 2012 with astounding results. The pass rate increased by 27%, and the course material retention increased from 9% to 91%.
Adoption has been slower for corporate learning. In a 2018 study by HR.com, 75% of respondents said their organizations view personalized learning as important, yet only 22% said they successfully provide personalized learning content.
Why the gap?
The difference between traditional and AI-driven adaptive learning
The techniques haven't become widespread in business until now because the individual attention a classroom teacher can give each learner isn't scalable. We need the help of our modern intelligent machines.
Building adaptive learning was a tedious chore before AI became a defining force in human capital management technology. It required a detailed analysis of the learner's current skills and performance to meet each learner's unique needs and interests. Developing assessments and decision points using algorithms to guide learning requires the development of detailed learning objectives matched to the content.
So, what trainers did when they needed to figure out what each person needed to learn was to start by doing assessments and putting people into groups. But you don't have personalized learning paths when you put people into groups.
For most organizations, the biggest obstacle has been the vast amounts of data required and the expertise to build heuristic models that can analyze the data in real time. When you use algorithms to manage learning, you must program every decision, but heuristics finds a practical, short-term solution. It's like our ability to use mental shortcuts by deciding based on prior experience. You must tell algorithms exactly how to decide and what to do; you can direct a heuristics model to find a solution, but it may not be the best.
And ... we're no longer impressed with gigabytes of data. The installed base of data in the world reached 6.7 zettabytes in 2020 and is growing 19.2% per year.
And most workers are becoming accustomed to using generative AI.
AI and our quest for personalized learning have led us to develop some cool tools. We can do assessments at any point in online learning, giving us valuable information to create specific learning paths based on how well we do in each module. We can mix and match all these bite-sized learning modules based on the assessment results.
The obstacles to making this work for most organizations have been the capacity to handle vast amounts of data that multiple assessments create and the expertise to build the algorithms and heuristic models to analyze the data and adapt the learning path.
But today, we're no longer impressed with gigabytes. The installed base of data in the world reached 6.7 zettabytes in 2020 and is growing 19.2% per year. (A zettabyte is 270 bytes, or one sextillion bytes, or about 1 billion terabytes.)
And, most workers are becoming accustomed to using generative AI.
The learning concepts that drive adaptive learning aren't new. The techniques you will use to guide your adaptive engine. Primary, secondary, and higher education practitioners have used them for several years. The principles of learning haven't changed … only the tools.
AI-driven learning hasn't become widespread in business until now because an individual instructor doing it manually isn't scalable. We need the help of our modern intelligent machines.
Nearly all the learning methods and theories are the same ones we have been using over the past few years to make learning more accessible and useful. They'll be familiar to today's learning professionals.
- Access to learning-on-demand in small bites that contain the information the learner needs at that moment to accomplish tasks.
- Spaced repetition to defeat the forgetting curve. We can predict when concepts will fall out of short-term memory to time the reinforcement.
- Gamification to engage learners and provide a sense of accomplishment using the motivational techniques developed by game designers.
- Allowing learners to select the modality they believe will be most effective. The options must be in context. Providing people with an array of choices of what and how to learn isn't helpful unless each person knows how he or she learns. For this function, we rely on metacognitive learning theory.
Metacognition in Learning
The broad definition of metacognition is thinking about one's thinking, but it's more than that: it's having control and awareness over our thoughts. [i] Metacognitive knowledge is everything you come to believe about yourself and others as cognitive processors, including how you learn, what you know, and what you don't know. Metacognitive experiences happen to us when we learn something new, solve a problem, or understand a situation.
These phenomena determine how we, as individuals, approach tasks and how we strategize accomplishing them. They influence everything we do and the "self-talk" we engage in when we're learning. They influence us to become hyper-focused on a task, abandon it, or just slog through it.
Many people who join our organizations haven't yet learned how to learn. Students trained in metacognition perform better than those who aren't. We recommend that every enterprise start each learner's development path with training in metacognition and structure learning to support each person's perceptions of their knowledge.
How Adaptive Learning Works
You are no doubt familiar with the concept of microlearning and the "chunking" of learning content to make it consumable on the go. Most learning organizations are well down the path to modularizing their content for on-demand mobile delivery.
Break down learning objectives into micro-objectives.
Adaptive learning requires breaking down content and assessment to a much more detailed level. Break down learning objectives into micro-objectives. A module with two or three objectives may become 40 or 50. When you break learning down to that level, it's obvious why you need AI to sort it all out.
Train your adaptive AI engine on your goal structure
A learning module usually begins with an assessment to learn what the individual knows and does not know. The individual should receive feedback on the findings and insights into how they learn best.
Your algorithms monitor each learner's progress and adjust the pace automatically, so learners spend their efforts on what they need to learn and don't waste time on what they know. By sensing when learners can't answer too many questions, the algorithms can provide a series of questions the learner will know to build confidence and then move on to more challenges.
The result is that learners know what they know and have confidence they can apply it on the job. They achieve true mastery instead of book knowledge they will quickly forget.
The Promise of Adaptive Learning
When your people can learn at their own pace and master the knowledge they need on the job, they will appreciate learning and be more willing to engage in it. Since they aren't re-learning what they already know, they will waste less time in learning activities and spend more time in productive activities.
Your new employees will become competent sooner, reducing the cost of onboarding. They will be more confident they can perform and will take risks to achieve more.
Your organization will reap the rewards of higher engagement and productivity when you can reap the rewards of truly personalized learning. You can:
- Personalize learning on a massive scale. You can train hundreds or thousands of people and give each one of them an individualized, personalized path to mastery.
- Adapt learning to the specific strengths and limitations of every individual.
- Provide content that engages learners by adjusting the pace to their level of confidence, challenging them enough to keep them motivated.
- Increase the net productivity of your organization while providing training far superior to traditional corporate learning methods.
- Provide an easily calculated, verifiable return on investment.
You won't become an expert in adaptive learning overnight. Each success will engender more success, and over time, your improved learning initiatives can create an organization that can withstand the storms of change.
How Pixentia can help
Pixentia is a technology solutions and support company dedicated to meeting the unique business needs of each of our clients. Our experts have over two decades of experience in learning and analytics, including the development and deployment of custom learning solutions that create measurable business value.
Pixentia is a full-service technology company dedicated to helping clients solve business problems, improve the capability of their people, and achieve better results.