On December 9, The Atlantic published an article entitled ‘The End of High-School English’ which warned that AI chatbots, like OpenAI’s ChatGBT, were already able to produce coherent responses to a variety of complex writing prompts. Immediately I went to OpenAI’s website, opened an account, and began playing around with the beta version of their text completion tool.
In the prompt box, I typed in a statement related to the writing assignment that I would be assigning to my students in January:
The lives of indigenous people in the Americas were devastated by contact with European explorers both immediately after contact and all the way to today.
The chatbot typed out this response:
The response, while articulate, was light on details and did not include any quotes or statistics to support the ideas presented. Nevertheless, I thought about how this tool would help some of my less motivated students produce serviceable answers to the many short-answer questions I used regularly in my classes.
What was particularly concerning about the text completion tool was the ‘Regenerate response’ button at the bottom of the page. When clicked, the chatbot produced a completely new response to the prompt, organized in a different way than the original, and featuring different vocabulary.
So, every student in my class could acquire a unique response, without opening a thesaurus or having to move sentences around. And, there would be no need to worry about being caught for plagiarism – when I put a line from the response into a Google search, there was no match anywhere online.
I spent the next few days thinking about my teaching practice. I regularly used short written responses as a way to have students extend their thinking about the text, images, and video that we explored in class. The existence of chatbots certainly increased the importance of discussion in terms of checking understanding and building upon ideas.
I thought about ways in which I could convince students of the importance of practicing writing. That the skills developed during classwork had value in the future. I came up with analogies to help students see the folly of outsourcing this skill development. Imagine, I would say to them sagely, if you were to allow AI to go to the weight room on your behalf. Yes, you would avoid all the boring and repetitive exercise. But, I would ask in my profound a-ha moment, would you be any stronger afterwards?
Some students would get it. But, most would opt for chatbot help, especially those who saw school through the lens of grades.
So, what do we do?
Chatbots are already here. In a few short years their exponentially evolving offspring will be capable of effortlessly churning out pages and pages of sophisticated and articulate essays on any topic imaginable.
How do we help students develop their thinking and demonstrate knowledge and understanding in that rapidly approaching future?
Sure, we could rely on anti-cheating software products. We could force students to write their papers by hand and in our presence. We could even require students to ‘defend’ their work in front of a panel of teachers.
Or, we can fundamentally reimagine education.
Reimagination won’t be popular. The other options I listed involve minor tweaks to existing and (somewhat) successful systems.
Reimagination will be complicated, expensive, and likely very painful.
But, fundamental change happens whether we like it or not. Darwin’s On the Origin of the Species, the automobile, vaccines, the internet, and even coffee were all resisted when they arrived.
So, instead of fighting chatbots, how can we integrate this amazing technology into our instructional and assessment practices?
- We could have students assess chatbot-produced writing
- Students could challenge chatbots with more and more complicated prompts to see if they can be stumped
- Students could conduct research and find interesting connections and then employ chatbots to perform the mechanical and menial task of packaging the ideas together
Here are the results of one crazy prompt:
Imagine asking students to analyze this kind of stuff, assessing the accuracy of the chatbot’s efforts to mimic different styles. Talk about engaging!
Like it or not, AI is here to stay. We need to reimagine education and get ahead of the tsunami that will inevitably follow.