An Educator's Guide to ChatGPT: The chatbot that could change everything
Monday, December 26, 2022
ChatGPT…maybe you’ve heard of it, maybe you haven’t…but it’s a big story.
“The New Chatbots Could Change the World. Can You Trust Them?” - The New York Times
“ChatGPT Wrote My AP English Essay—and I Passed” - The Wall Street Journal
“Why We're All Obsessed With ChatGPT, a Mind-Blowing AI Chatbot” - CNET
And while ChatGPT has the potential to profoundly impact a number of industries in time (e.g. it can already write movie scripts, entire essays, and even some code) - its greatest impact is likely to be in education. It will take some time to playout, but teaching and learning will likely never be the same - and, as educators, we should be excited about that!
In the most oversimplified terms possible, ChatGPT can be described as a chatbot - or a basic messaging system that you can interact with in a conversational way. Imagine if you were having a conversation with the Google search engine - but instead of you entering in keywords and it spitting out lists of potential answers and websites, it just responds to you directly with no need to navigate anywhere.
In more formal terms, ChatGPT is what’s called a “large language model,” which has a complicated definition that has to do with probabilities and words and stuff. Basically, it’s a form of artificial intelligence where they’ve trained a program to have conversations and perform tasks using words. This bit from the NYT story I also found helpful:
“Five years ago, researchers at Google and labs like OpenAI started designing neural networks that analyzed enormous amounts of digital text, including books, Wikipedia articles, news stories and online chat logs. Scientists call them “large language models.” Identifying billions of distinct patterns in the way people connect words, numbers and symbols, these systems learned to generate text on their own.”
ChatGPT is the “large language model” developed by OpenAI. Google also has a model called LaMDA, but it has not been released to the public yet.
It was developed by OpenAI, a research and development company based in San Francisco, whose mission is “to ensure that artificial general intelligence benefits all of humanity.”
That’s the thing, you can ask ChatGPT to do some crazy things like:
What makes it so special is that not only can it answer complex questions or perform complex language tasks, but it can do so in as many different ways and perceived levels of understanding as the average person could.
But nothing beats seeing it in action - check out our intro video where I enter in a few different education-related queries.
Technically the answer is probably yes and no because of how you define “wrong.” But for our purposes, it is important to know that ChatGPT can be inaccurate. This is the first of a few areas for improvement in the current model.
Not everything ChatGPT says is factually correct. Because ChatGPT answers queries through a statistical method, it is not actually learning things about the world, it is just amazing at guessing what the right answers might be and how they might be said. With this in mind, a good rule of thumb is to stick to using it for things that you yourself have knowledge in (e.g. teaching and learning) or to things that you can personally verify (e.g. math problems).
ChatGPT also suffers from a few other challenges at the moment:
These are some significant gaps, which will be ironed out over time in all likelihood.
Fortunately, there are still productive ways to utilize ChatGPT in teaching today, as long as the right precautions are taken.
Even with its flaws, ChatGPT can be an incredible tool for teacher and leader productivity. But before we list some potential use cases, please take note of this general precaution:
Treat the responses you get from ChatGPT like you would anything else that you might find when searching the internet generally - scrutinize & verify using your own judgment and research.
Now, here are a few ways educators might end up using ChatGPT to boost productivity:
Even if you were able to reliably use ChatGPT for just one of those suggestions, you probably can already imagine how much time it might save you versus a typical session on Google (which is why even Google might be freaking out a bit).
You can try ChatGPT for free here
That is one of many big questions. Of course there are those who are already worried about cheating - and indeed there are instances of it too.
Given how easy it is for ChatGPT to answer questions and how convincing it sounds, it was inevitable that enterprising students would use it on assignments. But it’s also important not to indict students before they’ve truly committed a classroom crime. This is a revolutionary new tool that no one fully comprehends how best to use yet, but that doesn’t mean we should expect (or try to force) young people to stay away from it. Given how much the future will be different for them, we should probably be encouraging them to explore new tools like this, albeit in a responsible way.
And let’s be real - if you were back in 10th grade and had an essay due tomorrow that you hadn’t started, could you really imagine your younger self resisting using ChatGPT, even just to generate an outline or opening paragraph? I can already hear students ruefully (and rightfully) point out that “technology is supposed to make our lives easier, so why not let it?!”
In many ways, this is not significantly different from how the internet can be used today for similar purposes. I remember the times when Wikipedia was the scourge of teachers everywhere yelling “It’s not a reliable source!” Yet, over time, Wikipedia has become a very reliable source and students have learned to use it to find both relevant information and other sources (as mentioned already, ChatGPT lacks this ability, which is a drawback).
As always, it will come down to finding a happy medium and recentering ourselves on what education is actually meant to be about. Ultimately, it is not actually about scores and records. It is about giving young people the tools they need to thrive in the real world. Ignoring or outlawing new technology is never the way to do that.
Instead, this is a tremendous opportunity to introduce students to artificial intelligence, digital search & research to verify information, and see where they take it from there before rushing to judgments.
Fortunately, a tool has been released to help you double check student work. Similar to Turnitin.com which is used to detect plagiarism, this new tool is used to give you at least the probability that what was written was written by a human versus an AI (like ChatGPT).
Here is the link - https://huggingface.co/openai-detector/ - simply paste in student responses and the tool automatically tells you what probability it is that your student’s work was written by a human or by artificial intelligence.
Check out this quick video to see us briefly test this “OpenAI Detector.”
Throughout history, new technologies have arrived that have the potential to shift the way we learn, work & live - and ChatGPT very well might be one of those innovations.
History also tells us that efforts to suppress or outlaw new technologies is neither effective nor advised. Demanding that students steer clear of ChatGPT is a surefire way to ensure that many actually seek it out - they are kids after all :). And when students explore new technologies without guidance or training, that is when truly negative consequences can occur.
Be proactive and do what you know best - teach students about this new development, its strengths & weaknesses, and set some norms about how they should/should not utilize it in your current class setting. And don’t be afraid to use it yourself, albeit responsibly and with the recommended amount of caution as we laid out here.
Every day brings a new piece of the future into the present - and it’s exciting to be a part of!
Because when teachers come first, students win...and teachers stay.