On the first day of co-teaching their course “AI and the Future of Thinking” at the University of San Francisco, Mark Miller and Larry Ebert asked students to share their reflections on artificial intelligence and how they use it. The instructors — Miller a theologian, Ebert a systems engineer and management consultant who helps companies respond to technological change — expected unbridled AI enthusiasm. That’s not what they got.

The undergraduates, many of whom have experimented extensively with generative AI since the arrival of tools like OpenAI’s ChatGPT, Anthropic’s Claude and Google Gemini, expressed wide-ranging skepticism.

“That was super surprising,” said Miller, the department chair of theology and religious studies at the private Jesuit Catholic university. “I thought they were all going to say ‘I love it. Don’t ever take it away from me.’”

Instead, “at least three-quarters were pretty negative on AI.”

The students mentioned potential job losses, factual inaccuracies, cybersecurity breaches , AI companionship gone awry and the frustration they felt watching peers “cheat” using AI. Some noted the environmental impacts of the giant computing infrastructure needed to train and run AI systems.

But their negativity largely focused on the sense that using AI tools made them, in Ebert’s words, “dumber.”

“Someone compared it to junk food,” Miller said.

AI’s effect on human thought is already the subject of growing research — a widely cited study from researchers at the University of Pennsylvania’s Wharton School of Business earlier this year found evidence of people increasingly deferring to AI over their own critical thinking, a phenomenon the researchers called “cognitive surrender.”

Steven Shaw, one of the researchers behind the study, noted that because AI tools can complete a range of tasks quickly and competently, it’s tempting for students to associate the technology with good grades.

“As a result, it can be easy to slowly slip into cognitive surrender without noticing,” the assistant professor of marketing at King’s College London said in an interview. “What is less obvious is what might be sacrificed along the way: practicing independent reasoning, questioning an argument and developing expertise.”

Now, students who have come of age alongside rapidly evolving generative AI are confronting questions about cognition in the AI era firsthand in courses that go beyond teaching creative coding, prompt engineering and applications of the technology in business, accounting and other fields to examine its impact on the human experience.

Shaw sees such courses as “enormously important.”

“Universities are places that cultivate diversity of thought, intellectual independence and critical thinking,” he said. “Those goals are much more fundamental than simply teaching students how to use a specific AI tool.”

What Distinguishes Human Thought?

Miller and Ebert initially offered their course in the spring through USF’s Honors College, which requires students to meet a minimum GPA, and began teaching it again for the fall semester in late August.

The class explores the impact of AI on human thought, including critical thinking and creativity, by drawing on the ideas both of contemporary thinkers and foundational philosophers from centuries past — René Descartes, John Locke and Jean-Jacques Rousseau, the latter of whom asked whether advances in knowledge, science and the arts necessarily make human beings more virtuous.

Students also consider psychological, spiritual and theological viewpoints, including Buddhist figures and prominent 20th century Jesuit priest and writer Bernard Lonergan, who believed human knowing involves the distinct interplay of experiencing, understanding, judging and deciding.

The course — taught in a city at the center of AI development — doesn’t bifurcate the technology into savior or destroyer, all helpful or all harmful. And even as recent public debate about AI has focused on whether it could one day threaten humanity’s existence , it doesn’t offer clear answers. The point, Ebert says, is to wrestle, deeply and intentionally, with the questions AI raises.

“The shared goal was really opening the inquiry and purposely having many perspectives, many frameworks, that we could introduce and explore with students,” he said. “A big part of it is this belief really in an examined life and an examined use of technology.”

‘I Saw Myself Going To It Blindly’

A driving principle of the class is that understanding AI’s impact on human thought requires first exploring what constitutes human thought. Descartes famously said “I think, therefore I am,” suggesting that thinking is fundamental to human existence. What happens in our minds when we reason through a problem, formulate a thesis or learn something new? And how do these processes shape our identities?

Delving into the nature of human thought “was one of the most impactful parts of the course for me because it allowed me to step back and understand the origin of human thinking and what knowledge means,” said Gabe Baruzzi, a 21-year-old senior who took the class in the spring and often lingered afterward to talk further with the instructors.

The students in the team-taught class of 20 students come from a range of disciplines — nursing students, humanities majors, computer science and business majors.

Baruzzi, who’s studying business analytics with a minor in environmental science, took the course in the spring because he wanted to reevaluate AI’s place in his life.

“I would notice myself opening ChatGPT at moments where I was too lazy to think, just using what was easiest instead of actually using my brain to think problems out or methodically plan tasks,” he wrote in his reflection at the beginning of class.

“Once I saw myself going to it blindly at times, questions started popping and racing in my mind,” he added in an interview. “Creative thought has always been something very, very important to me.”

Not The Only Course Like It

The USF course isn’t alone in examining AI through the lens of human thought and experience.

The University of North Carolina at Chapel Hill, for example, offers philosophy courses on AI and the future of humanity. In an Arizona State University class called “Does Generative AI Have Philosophical Wisdom?” students engage large language models on topics like free will, consciousness and skepticism to try to answer that question.

Similarly to USF, the University of Chicago last fall offered an undergraduate course called “The Philosophy of Artificial Intelligence: Mind and Model” that explored the nature of human thought and what AI can teach us about it — Descartes also appeared on its syllabus. About 75 students enrolled.

Benjamin Callard, a University of Chicago philosophy professor who co-taught that course with fellow philosophy professor Jason Bridges, sees college philosophy classes on AI as places to ask, and try to answer, “upstream” questions: not just how we use the technology, but what using it means.

“Knowing what something can do (whether AI or a car or a gene therapy tool or anything else) does not tell you what to do with it,” Callard said over email. “That question is ethical, rather than scientific, and there is no point in proceeding with a technology if we lack a good answer to the ethical question.”

Charish Holland, a 19-year-old double major in politics and economics who plans to study law, enrolled in USF’s course in the spring, at least in part, out of concern for how AI will impact her 13-year-old sister.

“I feel like she will lose her own individual thoughts,” Holland, a USF sophomore, said in an interview. Holland assumes such widespread AI use among her peers that when her friends relay information, she often feels the need to verify it.

“I tend to look it up myself, just to double-check whether or not what I’m hearing is factual,” she said.

Using AI Differently After The Course

Both Baruzzi and Holland acknowledge that the USF course may be self-selecting, naturally drawing students who already question AI more vigorously than their peers. Even so, signs of wariness have emerged elsewhere among young adults. A Pew Research Center study conducted earlier this year found that 55% of adults under 30 now say they’re more concerned than excited about AI. Graduates at commencement ceremonies around the country this spring booed and jeered speakers who mentioned AI.

At the end of the USF course, “I would say that most of the students were still kind of down on AI, but in a more nuanced and deepened way,” Ebert said.

Baruzzi has been using AI differently since he completed the class, relying on it more for technical information than for generating ideas, he said. He still turns to it to supplement his thinking, sometimes feeding an LLM an original idea and asking how he might build on it and what other questions he should consider.

“If it can help me with that while still not taking away from my original thought, I think it could be a good tool,” he said.

Both he and Holland emerged from the course with more questions than answers, “which I’m actually very grateful for,” Holland said. “I didn’t want a bunch of answers because AI is like a constantly changing thing.”