Optimizing Language Models for Dialogue
OpenAI is an AI research and deployment company. Our mission is to ensure that artificial general intelligence benefits all of humanity. We are excited to introduce ChatGPT to get users’ feedback and learn about its strengths and weaknesses. During the research preview, usage of ChatGPT is free. Try it now at chat.openai.com.
Great promises from the makers of OpenAI. The question is if anyone knows what the consequences of this will be in school. The answer is, of course, no. We do not. But it is inevitable what is happening right now, so we need to understand the consequences and where it will lead us. The discussions about closing access to the internet might result in us being sidetracked, to be honest. But of course, we need to discuss that too. This is a brand new world; should we embrace it or deny it? Is there a middle way here?
I just read this article by John Spencer, No, Artificial Intelligence Won’t Destroy High School English (Or Any Other Subject), and it was reassuring to read his thoughts about this.
It’s true that AI can already replace the high school English essay. But the goal of an essay isn’t merely to write a great essay. It’s the fact that the act of writing is a critical part of how we make meaning. We learn through writing.
An AI can take a complex informational text and distill it down to a series of notes in history class. But a hand drawn sketch note helps create the synaptic connections needed to move the information from short term to long term memory. You become a better conceptual thinker when you don’t use AI for note-taking. If we look at this diagram of information processing, we need students to get information into their long-term memory:
So, if a student is learning to code, that might seem obsolete. Why have a programming class if we won’t even have traditional coders in the future? But the process of coding will help her learn how the programming language works. Over time, she’ll learn how to outsource the coding to the AI. She might double-check her work with AI. She might ask the AI for help with certain questions or ideas. She might even start with the AI code and then edit it to make it her own.
A digital artist might ask the AI to do five different pictures and then he uses that as an inspiration for his own work. He might take two different sample images and mash them up in a sort of collage art. He might turn the AI off completely and work on something from scratch and then later try digital modeling just to see the difference between the two approaches.
In math, a student might attempt a problem and then use Photo Math to doublecheck their work. They might take handwritten notes and ask questions aloud as the teacher models how to find the p-value. But later, that same student might use the ChatGPT to ask a series of questions like, “What is the p-value mean?” Or even “explain p-value to a 15-year-old.” Or maybe a clarifying question like, “Can a p-value be used in a correlational study?” or even “How do I find the p-value using a spreadsheet?”