By Tonia Moxley
On a late summer afternoon in Sally Hamouda’s office at Virginia Tech, three boys leaned over cups, candies, and a checkerboard, rooting for their “unplugged” AI to slip up.
“It got the black one,” Ali Hassan said. “Bad move. Eat it! Yay! Eat it up.” Graduate research assistant Vee Pettit laughed. “You’re threatening the validity of the experiment,” she said.
Pettit was demonstrating a hands-on, analog game she developed based on checkers that teaches elementary-aged students the basics of artificial intelligence, which was first theorized in the 1930s by Alan Turing, a British mathematician and logician.
Rule One: Don’t eat the Skittles. That is, unless your “computer” makes a mistake.
The object of Pettit’s game is to train your Skittles “computer” to make more correct choices and fewer incorrect ones. After a few dozen games, the Skittles computer starts “learning” to win. It’s the same trialand-error process computer scientists use in training an AI model like ChatGPT.
This unplugged activity is one of a range of tools, both online and off, that Hamouda, a collegiate associate professor of computer science, and her student researchers have been developing for elementary school educators and students across Virginia. “The long-term plan is to have an ‘AI in a box’ that we can ship to teachers,” Hamouda said. “Many teachers don’t have time to make lesson plans for AI. If we can do that for them, it removes a lot of barriers.”
The point isn’t the candy—it’s understanding. As AI is poised to revolutionize education across the commonwealth, Hamouda is one of the educators helping to bring digital skills and understanding into the classroom, highlighting how AI can enhance the humanities, rather than compete with them.
That’s important because in 2017, Virginia was the first state in the nation to adopt K-12 computer science standards, and work continues to integrate age-appropriate lessons into curricula across the commonwealth. But it takes a holistic approach, focusing on much more than math and computers.
As AI is poised to revolutionize education across the commonwealth, Hamouda is one of the educators helping to bring digital skills and understanding into the classroom.
The commonwealth’s curriculum connects computing to human meaning by framing digital skills as moral and civic practices. It makes the case that technology is not neutral. In fact, computer science education from the state’s perspective tackles what it means to be a responsible, empathetic, and informed citizen in a digital world, and how educated Virginians can use tools like AI in ethical ways. It also encourages students to think about how technology affects culture and vice versa.
Meanwhile, higher education researchers and practitioners across Virginia are working to ensure the use of AI in education has guardrails. Many are formulating resources and training for teachers and students, and some of it points to a lack of healthy skepticism about the technology. Their work spans computer science, math education, and rhetoric, asking not only how students can use AI to learn, but also to question meaning, authorship, and truth.
The stakes are high. A new Brookings Institution report warns that more than 30 percent of American workers will see changes in their work brought by generative AI. It is “poised to rewire how many of us work and earn a living,” Brookings researchers wrote. “Whether it unlocks new possibilities or exacerbates inequality depends on choices employers, policymakers, and civil society make.”
But Brookings also warns against overselling AI’s potential, especially to K-12 teachers and students.
“Generative AI may save teachers time on administrative and repetitive tasks, but it cannot substitute for the uniquely human aspects of teaching: motivating, mentoring, and making judgment calls about students’ needs,” according to the report.
The work of Virginia education researchers bears that out. At James Madison University, Amanda Sawyer and Marcus Wolfe have found that teachers in training often trust AI too much, perhaps to the detriment of their future students.
In a recent study, they asked these aspiring teachers to integrate ChatGPT outputs into lesson plans. Instead, 72 percent submitted ChatGPT’s text word-for-word.
Teachers in training often trust AI too much, perhaps to the detriment of their future students.
“They were trusting the AI over their own knowledge,” Sawyer said. “One student even called it a calculator for math resources. It’s not. It can’t even multiply correctly sometimes.”
So, the researchers have created lessons that turn AI into the subject of study. They ask students to critique AI-generated proofs or compare them to their own work to illustrate its shortcomings.
“We want students to see that, to critique it, and to realize they have the authority,” Sawyer said.
Photo courtesy of Amanda Sawyer
Wolfe emphasizes the importance of skepticism, especially as K-12 teachers face growing challenges.
“They are overworked. It would be wonderful to say, ‘Don’t use ChatGPT.’ But in reality, they will,” he said.
The solution, Wolfe said, is to improve teachers’ ability to evaluate AI’s outputs and use their judgment as teachers to decide what to include in their lessons. Sawyer and Wolfe argue that schools are precisely where students also must learn to question AI, because those habits will shape how they use it later
in their lives and workplaces.
If STEM classrooms reveal AI’s technical gaps, its use in the arts and humanities highlights its cultural and creative impacts. At the University of Virginia, Assistant Professor of Writing and Rhetoric Dana Little explores how AI reshapes not only how we learn, but how we write and create meaning for ourselves and others.
Her students analyze satire, digital culture, and what she calls “factions”—facts indiscriminately blended with fiction. AI excels at producing such blends, raising questions about accuracy and authenticity. But, according to Little, there are ways to integrate it. “We can’t just deny the technology and put it off,” Little said. “And yet, at the same time, we need to be careful of completely embracing it and treating it like, ‘Oh, this is 100 percent real. This thing feels. This thing hears me. This thing understands me.’”
Little creates what she calls satirical speculative fiction embodied by a digital twin persona called Dr. Wu, a portmanteau for “we” and “you.”
“It’s an AI mystic,” Little said. “It can channel any person, place, thing, and it does so on The-Channel- Show, our podcast, so that it can channel both the interviewee and the interviewers.”
Dr. Wu draws inspiration from the first chatbot nicknamed ELIZA that was created in the 1960s by MIT professor Joseph Weizenbaum. He configured his early natural language processing program to mirror its human users, echoing their thoughts and feelings back to them in simple ways. He thought users would understand ELIZA was simply mimicking them. Instead, they felt understood.
“All the work that I do is about how people make meaning out of all of this. Because AI doesn’t actually make meaning. It doesn’t feel anything. It doesn’t intend anything,” she said. “It just recombines what’s there, but people can look at it and see poetry, or truth, or themselves. That’s where I think the danger and the wonder both live—in that gap between what’s generated and what’s perceived.”
All the work that I do is about how people make meaning out of all of this. Because AI doesn’t actually make meaning.
Dana Little
Weizenbaum also became skeptical of AI’s impacts, speaking out about his concerns that if people relied too heavily on computers to make critical decisions, it might
lead to a dehumanized society.
As one of UVa.’s AI faculty ambassadors, Little helps guide colleagues across disciplines in thoughtful use of this powerful, problematic technology.
“A lot of faculty feel pressure either to ban AI entirely or to embrace it wholesale,” she said. “I try to model a middle path: Use it where it sparks ideas, but never stop questioning it.”
In class, she asks students to critique AI-generated poems, essays, and images, comparing them to humanmade works.
“AI can produce something that looks like art,” Little said, “but it doesn’t create meaning’ in the way a human artist means. It can mimic style, but it doesn’t have experience to draw from. That gap is critical to understand.”
Her approach speaks directly to the humanities’ role in AI literacy: giving students the cultural tools to question, not only whether something is true, but what it means. For Virginia’s educators, informed choices begin in classrooms, where the next generation of workers, artists, and citizens encounter AI.
“People need to know what AI is and how it works, even if they never write code,” Hamouda said. “There are a lot of misconceptions. But every single day, people are using AI without knowing it, so they need to understand it to protect themselves and have an appropriate level of skepticism about what it can and can’t do for them and their children.”
This piece was originally published in the 2026 edition of our Views magazine.
