The conversation around academic integrity may be eclipsing the real value of AI in education.
Marcel Proust annoyed readers everywhere when, in 1912, he published a 13-volume novel that holds the Guinness World Record for the longest ever written. It would take the average reader more than 100 hours to finish In Search of Lost Time. Yet, these days it's possible to write a 1500-word, five-paragraph essay about the opus in under a minute (it took Claude 39 seconds to write this) without even cracking the cover.
I was in the classroom in the fall of 2022 when Sam Altman first became a household name alongside the release of OpenAI’s ChatGPT. I’ll be honest; in what was the first year that felt truly “post-COVID,” some of us educators didn’t initially appreciate the seismic shift occurring under our feet. My district scrambled to block ChatGPT from the servers, and students seemed a little slow on the uptake. But it wasn’t long before my colleagues and I were sitting in department meetings, wondering aloud what this meant for us, our careers, and our students.
Are students using AI to cheat?
Tech writer Kevin Roose attempted to provide concerned educators some guidance in a 2023 piece for The New York Times.
“I encourage educators—especially in high schools and colleges—to assume that 100 percent of their students are using ChatGPT and other generative AI tools on every assignment, in every subject, unless they’re being physically supervised inside a school building,” he wrote.
At the time, I hoped Roose was overly cynical. But nearly two years and countless anecdotes—including this recent one by Clay Shirky—have proven him more right than wrong.
Your students are likely cheating with AI. It’s a hard pill to swallow, but a necessary one if we’re to move forward. Instead of bemoaning the issue, it’s important to be clear-eyed about it. Because only then can we move on to the most pressing issue:
What now?
What are we really teaching?
It’s often said, but it bears repeating: so many of our students will end up in careers that don’t yet exist. They’ll need a much different skill set to succeed in those undefined positions than the one we’re currently providing.
So while teachers are rightfully concerned about AI’s impact on academic integrity, it’s worth asking: the academic integrity of which skills? And is it perhaps time to rethink them?
There were a few years of my career when I picked up a social studies prep or two, and being able to write a compelling response to a Document-Based Question (DBQ) was an important skill to teach my history students. But what happens now that they can simply ask AI to analyze sources and write their essays?
Effectively answering this question requires a thought experiment: What was I really trying to teach them with this exercise? While analytical thinking and writing skills come to mind, the underlying importance of a DBQ was, at least in my view, source evaluation. Beyond the essay organization and the proper citation method, I cared that students understood things like the validity of a source or the perspective of its author. And as we enter the age of AI and anonymous authorship, I think those particular skills may be more critical than ever.
So maybe today I’d keep my DBQs just as they are, and maybe I wouldn’t. Maybe the exercise evolves into a task that AI can’t complete in five seconds. Maybe it becomes a joint venture wherein the student collaborates with AI to evaluate a source or debate an author’s bias. Maybe I’d ask AI to create a fake primary source and have the students evaluate it for credibility.
In rethinking an assignment’s underlying skillset, I can begin to reimagine what quality education means in both my classroom and my students’ tasks. Instead of trying to catch students cheating with AI, how can we help them learn with it?
Rethinking assignments in the post-AI classroom
While certainly not an exhaustive list, here are a few other examples that borrow from my rethinking of DBQs:
- Mathematics: Instead of hoping students don’t use AI to cheat on, say, word problems, math teachers should recognize the core skills that word problems are meant to exercise. Skills like mathematical modeling, identifying relevant information in a text (reading!), and breaking down complex situations can be done differently.
Maybe students write their own word problems and use AI as an editor to check their thinking. Maybe we leverage AI to generate in-class simulations to help students acquire and apply these same skills in new and interesting ways. Maybe we rely more on opportunities for cross-disciplinary collaboration and bring in our art education colleagues to help design STEAM projects. Maybe we leverage AI as a cognitive tutor to recognize why kids are making mistakes and help them work through misunderstandings.
- Science: Instead of begging students not to have ChatGPT write their lab reports, science teachers should find alternative ways to boost student competencies in scientific methodology, data analysis, and drawing evidence-based conclusions.
Instead of following a traditional chronology that asks students to first run an experiment and then reflect on findings, have students read an AI-generated lab report that has clear errors in its data analysis or one that skips key steps of the scientific method. Have them establish the purpose of an experimentation process by understanding what “science” would look like without it.
- English Language Arts: Instead of relying on AI to churn out a five-paragraph essay in record time, English teachers should reflect on the underlying value of analytical essays. One of the core skills we require of students is thesis writing, which is not only an argumentation exercise but also requires nuanced thinking and decision-making.
While essay writing is an important practice, there are ways to exercise argumentative thinking in other, equally valuable ways. Maybe teachers leverage AI to create tasks that help students understand concepts like circular thinking, logical fallacies, and the impact of fake news. Maybe students are assessed on their ability to identify the strongest argument from a collection of op-eds based on the use and interpretation of sources.
They should absolutely continue to develop their writing skills. Still, in a world that’s increasingly “online,” maybe those writing skills are honed for things like attention-grabbing ad copy or “hooks” that will keep online readers engaged and scrolling. Maybe those writing skills appear in technical writing exercises and opportunities to establish a portfolio while still in secondary school. Maybe those writing skills turn into a first draft of a novel manuscript based on commercial sales or book search trends.
- World language: Google Translate is good now, and it will only get better. Instead of trying to catch students writing essays with it (it’s usually obvious because the essays are too correct), world language teachers should have AI act as a pen pal or conversation partner to bolster student language skills.
With AI chatbots, world language students can have multiple conversation partners in different dialects or social contexts who can provide immediate, gentle feedback. AI can even be instructed to make a cultural or linguistic faux pas for students to identify and correct, boosting their own sense of self-efficacy and confidence in the target language.
Not too many world language teachers use translation as a regular exercise in their classrooms (we see you, Latin teachers!), but AI could be used to enhance those tasks. Students could be instructed to work with AI to translate colloquial idioms or social media posts and evaluate whether the original meaning is preserved or lost. Or they could use AI to generate “impossible to translate” phrases and have students explain the cultural concepts that contextualize those features of the target language.
Reasonable concerns
Even though AI may present a once-in-a-generation opportunity to reimagine many aspects of instruction, some elements of the technology require circumspection.
A man-made tool, AI is far from the faultlessly accurate, unbiased “Ask Jeeves” descendant it’s made out to be in some circles. There are reasonable concerns over how it’s trained and its ability to be manipulated.
Alyson Klein pointed this out in a 2023 article when she noted: “Students also need to understand how biases in the data that’s used to train AI can allow the technology to continue to perpetuate discriminatory policies unless humans recognize the problem and do something about it.”
This is where teacher training becomes so important. Teachers need to understand how AI works so they can educate students and warn them of its undeniable pitfalls. Unfortunately, professional development around AI use isn’t where it should be, and educators are left to use their professional judgment and research skills to fill in the gaps.
Districts and schools should prioritize professional learning that highlights AI and devote time and resources to crafting clear AI policies to best meet the needs of students and teachers.
Embrace learning by doing
Teachers who are interested in using AI to reimagine education should keep one key thing in mind: perfect shouldn't be the enemy of good. Your AI-optimized task does not have to be faultless in order to be valuable. But if you do need guidance in making the most of AI, we’ve got some resources for you:
And, of course, hands-on learning! Teachers don’t need to adopt every possible use of AI overnight; instead, they should choose one or two aspects of their job or daily life that may benefit them and see what it can do.
A new kind of academic integrity
Ultimately, artificial intelligence is here to stay, and—in the words of friEDtechnology’s CEO, Amy Mayer—it’s “not something we can fully say ‘no’ to.” Spending too much time bemoaning its impact on cheating is time we could spend learning how to harness this incredible technological development for good in our industry.
The opportunity AI provides us to reimagine education marks the dawn of a new kind of academic integrity, one where we can dare to rethink what we’ve always done and instead chart a new path forward.
The challenge isn't just to adapt to AI; it's to use it to create something better than what came before.
