�鶹��Ѱ�����

Illustrations by Dana Heimes

“What are you doing about A.I.?”

Outside of where to find a good cup of coffee or what the Wi-Fi password is, there’s perhaps no question you’re more likely to hear on a campus than how��
generative artificial intelligence is changing the college experience.��

Mostly, you’ll hear how professors are teaching students to use it like a tool, not a magic trick, or to think critically about it. But that’s exactly what they said about Wikipedia a generation ago. What’s really happening around A.I.?��

We asked people across CMDI to tell us how they’re actually studying and using—or not—A.I., whether in teaching, practice, research or creativity. (And for fun, we asked each person to describe A.I. in one word.) Here are their stories.

‘A��dz������DZ�پ���������’

“Move fast and break things” is the mantra of the tech industry. In his A.I.-infused art and teaching, Christopher Carruth instead embraces what he calls “a slow,��contemplative resistance.”��

“My practice is about using technology to interrupt, interrogate and agitate conventional systems,”��said Carruth, an assistant teaching professor of��information science.

An example from his creative practice is , in which he took algorithmically curated digital spaces—Facebook feeds, recommender engines, inboxes—and used code to disrupt their underlying visual logic, rendering them as abstract shapes.��

“These systems are increasingly shaped by A.I. and��predictive algorithms to capture attention,” Carruth said. “This work interrupts that rhythm. Rather than extracting attention, it creates a space where viewers can become conscious of how their attention is being shaped in��the first place.”

Notably, his work is not about rejecting A.I.: In his classes, he encourages students to debate ethical issues—self-driving taxis, predictive policing, facial recognition—to get them to slow down, too. He enjoys when students disagree and challenge their own ideas about technology.��

“For a student who is focused on how to build a large language model, I hope to shift the thought process to how the data will be collected, and what consideration they have for copyright,” he said. “And having that dialogue without making it about right-vs.-wrong.”

‘P��������dz澱������’

Pooja Iyer likes experimenting with technology, so it’s no surprise she was among ChatGPT’s early adopters. “I remember immediately thinking this was search on steroids—so, personalized ads on steroids,” she said.��

Iyer, an assistant professor of advertising, studies the interplay of data, privacy and personalized marketing. As A.I. platforms embrace ads, it’s clear we’re not going back on the ad-supported model of the internet.��

“But how do we build this balance of being ethical and mindful about the data I, as an advertiser, use to sell my product?” Iyer said. “Because we’ve never asked the public what data they’re OK with advertisers using.”��

Her research is trying to answer that question. ��

“Can I use your demographic data to show you a��personalized ad—your behavioral data, health data, voting behavior data?” she said. “Research has��consistently found that people like personalized��ads, but there are ethical questions about how��deep in the data we should go.”

Iyer typically opens her classes with a project where��she doesn’t offer guidance on whether, or how, A.I.��can be used.��

“I wind up with 10 projects where I cannot tell one from the other,” she said. “And these students are competing for the same jobs. So, if this is your output, and it all looks the same, why should I hire you? Why not just plug it into A.I. myself?”

‼�����������ٱ��’

Ask Greta Kerkhoff (Jour’26) where the stands on A.I., and she’ll point you to a very concise, one-word statement forbidding staffers from using it��in their work.��

“We pretty much have a zero-tolerance policy,” Kerkhoff, who served as editor-in-chief of the CUI last year, said.��

Kerkhoff completed internships at KGNU and the Denver Gazette; she also published a report on�� in The��Colorado Sun late last year.��

She isn’t a Luddite—she recognizes these tools can be helpful for brainstorming, for instance. But she doesn’t buy the hype that it makes journalists more efficient.��

“Even if there was a way to ethically use it, it creates more work for reporters because you can’t trust it’s telling you what’s right,” she said. “And the models��are trained on existing information. If you rely on an algorithm to suggest sources, it will just regurgitate what’s already been done.”��

Before Kerkhoff became its editor, the CUI briefly published a piece that the staff eventually determined��to be written by A.I. It included multiple hallucinations and led to the dismissal of a writer and an editor.

“There are such ethical dilemmas around this��technology,” she said. “If we don’t set some parameters around it, we’re going to live in this��world of increasing digital murkiness.”��

‘U�Բ����Ǿ��岹������’

Chris Vargo calls himself a lifelong computer nerd��but someone who “barely spoke the language of the��country he was in”—until he discovered ChatGPT.

“Now, I can climb Mount Everest, with software,”��said Vargo, an associate professor of advertising��who specializes in analytics.��

Vargo has published plenty of papers that rely on A.I.��to uncover insights in datasets from Reddit, YouGov��and others. A co-authored with Toby��Hopp, also an associate professor of advertising,��and PhD student Shahed Rahman examined the different customer journeys of people who use large language models, or LLMs.

“This dataset is two years old, but even then, one-third of the people were using LLMs,” Vargo said. “Yet��advertisers haven’t updated their thinking beyond��the traditional customer journey funnel.”��

That funnel, which guides consumers from awareness to action, is collapsing—LLM users tend to make faster buying decisions, cutting out the middle—and business hasn’t responded.��

“Ten years ago, to generate leads, you put a white paper behind a gate and made people sign up with their email,” Vargo said. “But if the agent can’t access that information, it can’t suggest you to users.”��

He recommends companies put relevant information in easy-to-scrape text files—something he’s done with his own work, so A.I. agents will extend its reach.

“If marketers make it hard for agents to get the right data, they make it hard for their own products to be found,” Vargo said. ��

��

‘H�DZ��ڳܱ�’

Don’t think A.I. has its eyes on religion? Tell it to Jesus. Specifically, an A.I.-generated messiah that hears confessions in a Lucerne, Switzerland, church.��

“I don’t have all the answers, because it’s moving so fast,” said Nesrine Mansour, assistant professor of architecture. “But A.I. is affecting how people are interpreting religion and thinking about architecture, authority and spirituality.”

Mansour studies spaces in digital religion—like games, televised services and apps—and the interplay of��sacred architecture, religion and digital media. She’s developed guidelines for virtual sacred buildings to support architects, artists and game designers.

“I surveyed these virtual churches in game environments, and found most of them were replicas of real ones,” she said. “In a virtual realm, where they aren’t restricted by budget—or gravity—why do people��re-create real churches?”

A book she edited, Religion in the Age of Artificial Intelligence, explores how A.I. shapes our day-to-day and spiritual lives, so that we might engage with it more thoughtfully. And A.I. is key to her teaching: Mansour incorporated diffusion models and LLMs into classes when the technology was just months old, and constantly tinkers with how to use A.I. in class.��

“In architecture, we care a lot about how the building or rendering looks in the end,” she said. “But in terms of learning, I really think ethical workflow processes and prompt engineering deserve more attention.”

‘B��ɲ�����’

A.I. is not all bad, Mehak Sawhney tells her students, but nor should we be all in on the technology. Instead, she advises A.I. be small scale and for special purposes.

Sawhney, an assistant professor of media studies, studies the environmental—especially water—costs of A.I., surveillance and other topics.

“A.I. needs to be understood across scales: global, regional, national and hyperlocal,” she said. “It is only through a combination of these scales that a nuanced understanding of A.I.’s role in financial bubbles, its varied adoption and its environmental impact can��be understood.”��

It’s the sort of humanities-focused approach she’s brought to a new graduate course on A.I. and the environment that she’s teaching.

“From an ethnographic perspective, it is important��to see what A.I. means in different contexts,” Sawhney said.��

For instance, people in developing nations��might approach A.I. as an equalizing tool, especially��for non-English speakers, while those living under authoritarian rule might see it as a progressive��research field.��

“However, this points more toward the colonial��hegemony of English, ethnic and religious supremacy, and lack of resources in many parts of the world, rather than A.I.’s ability to solve structural problems,” she said.��

‘T�����Բ��ڴǰ������پ��DzԲ���’

By day, Priscilla Rain Hopper is a teacher who��specializes in A.I. By night, she is … also a teacher��who specializes in A.I.

Hopper (InfoSci’22; MS’23) is a lecturer of information science who works for an A.I. company—“which I said��I would never do,” she admitted—that provides teachers and K-12 school districts with tools that allow them��to more effectively plan lessons and provide support��to students.��

What drew her to MagicSchool AI, where she is a senior LLM quality analyst, “is that it’s A.I.��for good.”��

“My students want to��fight the bad,” Hopper said, noting their concerns about ethics, environmental impact and employment prospects. “But you can’t do that from the sidelines. So, my message to them is: If you believe the rest of the world has the same ideas, intentions and solutions as you, fine. But I don’t think that’s the reality; otherwise we wouldn’t be here.”��

At CMDI, her focus is teaching students to see the ways different A.I. tools can be used together to improve productivity—so, to be conductors of the orchestra,��instead of experts in a single instrument. She also created a website for students that prepares them to position themselves in a competitive job market where the A.I. skills core to the information science major can prove a differentiating factor.��

‘R���ܱ���������’

It’s not wholesale A.I. that upsets Lori Emerson. Rather, it’s how far we’ve landed from the idealism of the 1940s and ’50s, when A.I. was first theorized.��

“This shift we’re living through—where corporate A.I. is embedded into every piece of software and every platform—is not inevitable,” said Emerson, a media historian and professor of media studies. “This is being foisted on us by corporations trying to make as much money as possible.”��

Last year, Emerson published Other Networks: A Radical Technology Sourcebook, which pushes back on the idea of inevitability—“that every change rolled out by technology companies is for the best, and makes things easier and more efficient.” She challenges students to question the narrative that they will be irrelevant if they object to A.I., and encourages them to think about older, analog technologies—like books, pen and paper, and typewriters—that enable creativity and invention, without conceiving of the user as purely a worker striving for efficiency.��

“Partway through the semester, students begin to understand that thinking, writing, producing, researching on their own terms—without it being done for them—is maybe the heart of what it means to be a human,” Emerson said. “That’s why they come to college.”��

Joe Arney covers research and general news for the college.