Tailored guidance, real-time feedback, and scalable support for every learner is a reality. But as associations embrace these tools, it’s essential to weigh the benefits against emerging risks.
Wouldn’t it be great if we could have an all-knowing mentor — someone who knew how to perform every skill we needed to learn in whatever area was important to us? Someone who would watch over our behavior like an apprentice’s master, correcting our mistakes in real time based on their background and years of experience? Kind of like Neo learning martial arts in the Matrix by plugging his brain directly into the programming for karate, jujitsu, and taekwondo. This is truly learning at the point of need.
Wish no more! Those who work in association professional development now have access to tools that can provide many of these services to their members — tools that maximize their motivation for learning and the rewards they experience for it. According to ASAE’s 2024 ForesightWorks document “AI and Learning”, these “swiftly advancing artificial intelligence technologies, including generative AI, will transform how and what people learn, driving rapid change in education models and content, and affecting training and skill building.”
Two generative AI approaches that are separate AI behaviors but often work in tandem are “adaptive learning” and “personalized learning.” Here’s how they were implemented at an association we’ll call “The Association of Big Truck Dealerships” (ABTD) to quickly upskill its novice truck dealership members.
When “Phil,” a senior education specialist, was charged with seeing what AI can do and specifically told to determine which of 150+ LMS offerings would best serve their membership, he recommended a platform designed to deliver learning at the point of need with adaptive and personalized learning experiences. This was important to enable ABTD’s beginning sales associate members to quickly get up to speed on how big trucks operate, as well as how to be effective in sales. He explained to leadership how the AI tracks learner progress and generates more activities and assessment items to help users improve in areas where they are slower to learn or selecting incorrect feedback. The LMS also had an intelligent chatbot that new sales associates could use to ask questions in a conversational manner and receive an accurate response with explanations and guidance.
[These] swiftly advancing artificial intelligence technologies, including generative AI, will transform how and what people learn, driving rapid change in education models and content and affecting training and skill building."
AI-powered “adaptive learning” responds to learner behaviors in real time by adjusting task difficulty, recommending resources, and delivering tailored feedback. In “observing” learner behavior, this form of AI assesses, for example, how the learner is progressing, the length of time they are taking to respond to assessment items, and what choices the learner makes. It then compares learner behavior against complex algorithms it has acquired through a massive input of information, which determine statistically, “if the learner does this + that + the other, then the learner needs more to help them process the information or even perform another application activity. It then produces real-time, individualized feedback based on what it has statistically determined is required at the specific point of learner need.
“Personalized learning” considers member learning styles, such as being a visual, auditory, or kinesthetic learner, and personalizes recommendations for additional training based on “observed” learner behavior. Does the learner tend to choose video resources to understand material? If yes, the AI will identify from its immense database the specific YouTube video that statistically aligns with that learner’s pathway at that point. It will then instruct the learner to “go to minute 36 and watch the next 11.5 minutes, ending at minute 47.5” to gain a greater understanding of the concept being learned. If the learner is statistically behaving as a small specific subset of learners in its model that requires additional content, it might create a PowerPoint, replete with colorful images and bullet points that engage the learner. It may also deliver it as a PDF that the user can download and print — because part of their learning style is to write directly in the document while they are learning.
Personalized learning AI works hand in hand with adaptive learning engines to monitor student progress, intervening when necessary to ensure learners stay on track. Based on learner behavior, AI-powered systems and platforms can provide personalized guidance, support, and feedback, catering to each student’s specific needs. Whereas a mentor can guide one or a few individuals, the power of adaptive and personalized learning resides in its ability to immediately respond to hundreds, thousands, or even more.
“AI-powered systems and platforms can provide personalized guidance, support, and feedback, catering to each student’s specific needs.”
Sounds pretty fantastic, doesn’t it? Back in the “day” — no more than a few decades ago — these AI processes were established through the development of algorithms, which were trained with “if this, then that” or “if these, then not those.” As AI engines grew more powerful and were fed massive quantities of (hopefully) relevant content, AI algorithms were able to incorporate models of how experts and novices — and all learners in between — performed behaviors, what knowledge they needed to perform those behaviors, and what the typical and not-so-typical pathways learners took to learn new skills.
By coupling the two customized learner support tools with the following two additional AI processes, knowledge-seekers are fairly guaranteed to be hooked in their learning journeys. They will feel empowered, engage more deeply in the content they are learning and — with the provision of tailored assessments and corrective feedback in real time — be motivated toward greater mastery of the material.
The two additional AI processes are:
With that understanding in mind, let’s return to Phil and his new learning management system. It’s been up and running for the last year. ABTD’s members are overwhelmingly positive about how much they have been able to learn and retain so quickly, as well as transfer over to their jobs. Behind the scenes, the AI generates a unique learning path for each sales associate through “observing” learner behavior. If, for example, one learner likes to read and absorb information from podcasts and webinars delivered by different truck manufacturers on “spec’ing” a truck, the AI scans the internet and identifies the right content in the blink of an eye. Since it’s only a subsection of the content available, the AI directs the learner to listen to those specific points or watch the relevant discussions. If another learner likes video and wants to learn more about the nuts and bolts of selling a truck, the AI will find videos that include social learning models so they can learn by watching others.
These four processes, then, are the rapidly developing approaches that will scale to thousands or more — simultaneously. And there are many products out there, such as Yourika, Canvas, Realizeit, Absorb, and Brightspace LeaP by D2L — with many more existing or in the pipeline. Together, these approaches can create in learners a hunger to learn more. They appear to be founded on the tenets of instructional motivation, as discussed in Keller’s ARCS model:
And herein lies the cautionary concerns.
Much of the discussion about these products and approaches currently centers on the benefits, and certainly there are many. But there are downsides to this age of personalized and adaptive learning.
The first potential challenge is a concern about the psychological impact these experiences may be having on learners. Because of the tailored, seemingly authentic, “give and take” interaction with the machine, learners may psychologically attach to their AI support because it seems “natural.” They are our “friend.” But AI doesn’t fondly remember experiences of mentor and mentee working together.
AI adaptive and personalized learning approaches are based on a massive trained model of how thoughts and words statistically go together regarding a specific skill or knowledge set. It compares them to countless models of knowledge to determine what, based on probability, it should do next. Ethically speaking, are we building a generation of learners who have come to rely on a mathematical engine as they might rely upon a human mentor?
There should also be concern about the reward structures provided to learners as they become increasingly engaged in learning because they are being rewarded for behaviors as drawn out by AI. The more we get rewarded in this continuous learning culture, the more we will want. The brain delivers dopamine in response to rewards. More rewards mean more dopamine. The question should be asked — even if it is a bit of a stretch — are we potentially creating a generation of addicts that depend on the machine for their “fix”?
Lastly, there is value in the struggle to learn. Our brain structures develop and strengthen through trial and error. We are programmed to figure things out. If some or all of the learning struggle is removed through these AI interventions, the question to ask is: What fundamental brain processes and linkages are changing as the result of these interactions, and what will be the result in our behaviors going forward? Will we continue to have the chance to be resilient through our struggles, or will this incredible tailored learning experience lead to dependence?
Today is the age of the AI algorithm — and the approaches are improving literally minute by minute. The opportunity for associations to deliver high-quality individualized learning experiences will give members more power and ability to advance in their careers. This in turn will make associations even more of a go-to resource for member professional development. As this trend will inevitably go forward, may we balance the many amazing affordances that adaptive and personalized AI clearly deliver to learners with a healthy dose of caution and care.
