5 thoughts on how AI will be used to teach humans and what the problems are
I’ve been an adult educator and vocational trainer since 2016, and I incorporated GenAI + GPT tools into my curriculum in 2021, so I have some insight here.
1. “Give Me Examples”
The first thing most people use AI for is ideation. As a creative tool (someone might say ”creative partner”) AI is great, because it relies on the human’s discernment and taste to shape the final output. This can cause problems if the human just takes the AI’s response and copy/pastes it without transforming it. You’ll see this theme a lot from me…
2. “Do: Task”
Instructing the AI to perform a task is another thing people do right away. 99% of the time, they ask the AI to do too big of a task, with far too little context. This generates an extremely generic and often wrong output.
Instructing the AI to do very specific tasks with lots of context results in great, often creative and helpful outputs. The trouble is that reaching this level of detail with the right context is hard and takes a while, so most avoid that effort.
3. “What is…?”
Asking the AI for factual, contemporary information is currently ineffective. It doesn’t have direct access to the internet, and when it does, it doesn’t “search” the same way we do.
Instead, it interprets your prompt, decides that it needs to search the internet, decides on a search query itself, takes the results and attempt to discern their meaning, and the generates a response based on its discernment. All of these steps are done without human insight or context, and they rely on the currently most prominent search result (which can be wrong or out of context)
4. “How do I”
Asking AI to instruct will result in pre-2021, most popular instruction. This may not be optimal.
Also, most learners don’t actually know what they need to be learning, and AI can’t tell the current level of the learner to guide them. So a learner might go down a rabbit hole that is unnecessary or unhelpful.
5. “Help me do X”
This is where AI shines: as a helper and assistant that isn’t required to have taste or make important decisions.
collaborating with AI to learn can tighten feedback loops on deliberate practice, and as long as the topics are well research and we’ll-known (like most K-12 curriculum) the AI has great knowledge here to help.