AI’s Emerging Role in Learning Experiences

Generative AI has exploded on the world’s radar with wide-ranging implications for society. We don’t yet fully know what to make of its power, let alone how it will impact the learning space. But we do know that in order to talk about it, we have to be speaking the same language. So to discuss AI and what it can do in learning spaces, let’s first define what we mean; then we can delve into the cutting-edge of AI in learning experiences.

Studion’s Senior Learning Experience Designer Marc Campasano offers the following definitions of generative AI and large language model (LLM), which are widely accepted within the industry: “Generative AI means AI that generates or creates new content. A large language model is generative AI that uses a large body of existing text to generate new text,” Marc says. “It’s creating an internal model of how words relate to each other, and it’s starting to categorize things and work with relationships and causation. This is what allows it to make new content.”

When it comes to an agreed-upon definition, the trickier term is actually AI. “Previously, almost any time a computer was simulating a person or character, we were calling it AI, but that definition doesn’t hack it in 2024,” Marc says. “I would say AI is when a computer learns a problem-solving capacity beyond those hard-coded by a creator. The creator gives it a goal and a ton of data, and the computer invents novel ways to achieve that goal. I say ‘goal-oriented’ because a lot of the early modern AI was on games like Chess and Go, which were fed thousands of games and the rules, but that was all that was given. Now, it’s just ‘Write an essay,’ and the computer hasn’t been told what an essay even is. The users tells it to go ahead, and it has this very adaptable framework that lets it solve the problems.” This adaptability is what allows AI to serve as an effective tool in the learning experience designer’s arsenal when putting together a course or building a chatbot.

Current AI Use in Learning Experience Design

Specifics about what these AI capabilities mean for learning experience design, and educators and learners themselves, are still revealing themselves. Three current examples give a taste of what’s possible:

• Kajabi, which provides an all-in-one business platform to help entrepreneurs create and sell their digital services and products, is experimenting with a new AI Creator Hub that can quickly develop course outlines using their generative AI tools.

• Duolingo Max is using AI to provide learners with personalized responses to their answers using two tools: Explain My Answer and Roleplay. With Explain My Answer, learners interact with a chatbot that tells them why their answer was right or wrong. With Roleplay, learners play out different scenarios with various characters, giving them a chance to practice real-life conversation skills with a trained, responsive AI. “These answers are vetted, tweaked, and monitored by humans, but they are being made up on the fly by the AI,” Marc says.

• Khan Academy’s digital tutor, Khanmigo, is similar to Explain My Answer, giving students individualized guidance and responses to questions and prompts they enter. It also provides teachers with help making lesson plans and rubrics and alerts them when students are struggling within Khan Academy.

These types of options provide learner-centered content, actively drawing each learner in by meeting them where they are.

At Studion, we and our clients are also leveraging generative AI. In some learning experiences, it’s used to assist the creation of new content in partnership with human editors. Marc, who has assisted clients with this work, says, “The best case for using ChatGPT to write content right now is for inspiration, the first draft, the structuring. It’s a tool to aid writing, not a wholesale replacement for a thoughtful human.”

In other learning experiences, AI has been used to construct project teams across interests and time zones, as well as to build rubrics and grading models. One of our clients has had success grading student essays this way. First, they use natural language processing models to build a grading model in-house, and then they fine-tune the model using essays that have been hand-graded by a person. This facilitates a faster, more cost-effective grading process.

Potential Future Trends in Learning with AI

The successes of tools like Duolingo’s Explain My Answer and Kajabi’s Creator Hub bode well for the future of generative AI in learning spaces, especially when cleverly used to assist a human creator. Starting with great prompts (also known as prompt engineering) is one key aspect of creating trustworthy, well-written, and engaging content with AI. As Marc says,

“If you want AI to write anything good, you need to understand the topic and the quality that you want. You need to have the chops to know what to ask and to know what good looks like.”

People must lay down the parameters of a learning experience because AI cannot vet its own content or rate its own quality.

There’s also the possibility that one day, chatbots will be more accurate because they’re specifically trained to be niche. Right now, a lot of the hallucinations (that is, the false information AI invents and presents as fact) and mistakes that arise from interactions with ChatGPT happen because it’s trained widely on many topics, but not trained in depth on any of them.

Marc’s favorite example of what a possible solution is this: “If you’re the Classical Music Museum, you feed your chatbot everything you trust about classical music, every fact and other baseline knowledge, and you train it to say ‘I don’t know’ when asked questions that are not about classical music. Then you’ve got a reliable expert on one thing. You’re not being constantly inundated with new information until you, the controller, choose to teach the chatbot new things. It becomes an expert based on that body of work.” This type of specialized chatbot could contribute to active learning opportunities for learners, keeping them motivated and interested and helping them achieve their goals.

As AI of all kinds continues to advance and the education industry reckons with the resulting changes, learning experience design will undoubtedly change as well. At Studion, we’re excited about expanding opportunities to leverage generative AI to get more impactful learning to learners. It’s all about making courses more responsive to keep people engaged, thereby improving learner retention.

We can also use AI to connect learners, whether because of their similarities or through their differences, so that they’ll support each other and build a sense of community. Together, this active engagement and connected community will lead to more meaningful real-world outcomes for everyone.