Many are looking to generative artificial intelligence to create a scalable, effective approach to delivering one-on-one tutoring to learners. As we’ve discussed in an earlier post, that is an exciting direction, to be sure, but there are many other potentially interesting applications when it comes to learning. In particular, content can be generated quickly and unlock other innovative modes of learning. Here are a couple of examples.
1. Simulations and scenario-based learning: creating real-world context
Across learning contexts, relevance matters. If a learner can’t address the question of “how will I use this?," they are likely to be unmotivated and not get a lot of value out of the learning experience. Generative AI can help by immersing learners in customized scenarios, tailored to an individual. Consider someone trying to develop their project management skills. With just a few guiding sentences in a prompt to ChatGPT, they can receive a tailored project-based learning scenario:
Someone who is more novice can be walked along the same learning path in a more structured, step-by-step way.
It can be kind of shocking to see how entire learning paths like this can be developed for any topic or skill in just a few minutes!
2. Dialogues and conversations: Enhancing communication skills
Many critical tasks we engage in involve some kind of communication, whether presenting a deliverable or negotiating a contract. These kinds of dialogues can be difficult to improve because they may happen infrequently or be so high-stress that negative emotions can make them difficult. Practicing these with a generative AI can be a great way to get in the repetitions necessary to develop skills. Researchers can try out an interview script to get a feel for it before meeting with a user or customer. Small business owners can pitch to a virtual Shark Tank to practice their presentation before seeking funding, and get personalized feedback. You can even ask for the responses to be tailored to sound more like Mark Cuban!
3. Teaching, not just tutoring: Explaining concepts to virtual students
As mentioned earlier, many organizations are trying to leverage generative AI to improve access to effective tutoring. But teaching a topic can be a very effective approach to learning. Rather than asking ChatGPT to teach you, consider reading up a bit and then trying to teach ChatGPT. With careful prompting, you can get ChatGPT to act as a simulated student, and the act of teaching will help you solidify your own understanding.
Things to be aware of when learning with generative AI
Of course, as with any use of generative AI, it is important to check what it produces. Hallucinations are possible, particularly in facts or citations it generates. The feedback the system generates may not be as accurate or valuable as from true experts, but those experts may not always be available to you, and certainly not as readily as the OpenAI or Bing websites. The access vs. quality trade-off is always going to be one to consider. And, of course, there can be social and emotional aspects of learning that can come from interacting with peers and experts you admire and relate to; it is important to consider these kinds of tools as supplements, rather than replacements, for learning from and with other people.
Finally, while it is surprisingly easy to get the AI systems to produce these kinds of learning activities, it does require some amount of patience and prompt engineering to refine the experience to match exactly what you are looking for.
