Interdisciplinary AI Ethics class launches Fall 2026

Will AI become superintelligent? Would superintelligent AI take care of us? Or would it become evil and kill us all? The professors who designed the University of Utah’s new AI Ethics class think that is only one of many worries—and some of the other ones are affecting us right now.

As artificial intelligence grows rapidly and becomes more embedded in the technologies and systems that affect us every day, U students wonder what it all means for their future—and this class has arrived at precisely the right time. Coming out of the One-U Responsible AI initiative and co-developed by associate professor of Philosophy C. Thi Nguyen and Kahlert School of Computing professor Jeff Philips, this class will help future students navigate ethical questions around AI design, regulation and use.

“Super intelligent evil AI? Maybe. AI wreaking climate havoc through water use? Definitely. That's happening. Job displacement is happening. Bias is happening,” Nguyen said.

The class addresses AI issues happening right now, such as fairness and discrimination in automated prediction, authorship and ownership in generative media, environmental costs, replacement of work and impacts on human flourishing. Classes will read case studies, papers by AI ethicists and engage in discourse surrounding these topics to guide their responsible creation of future AI products—because today’s students will be tomorrow’s engineers.

The new AI Ethics class PHIL 2150 / CS 2395 will be available for computer science majors who have taken CS 1410 or CS 1420 and for philosophy majors without a prerequisite. It’s also available as an elective for any undergraduate student and will be required for students in the new AI Major, which starts in the Fall semester of 2026. 

Embedded EthiCS and ‘aggressive interdisciplinarity’

One of the inspirations for the class was Harvard’s Embedded EthiCS initiative, created in response to students’ request for more focus on the moral results of technology. Ethics of computer science classes didn’t work well when only technical staff, computer science, or engineering faculty taught them, because they usually didn’t have ethics training. And when philosophers taught classes like this, the concepts were often too abstract to connect materially with engineering. Embedded EthiCS, a form of what Nguyen refers to as aggressive interdisciplinarity, grew out of the idea that this type of class only works when philosophers and computer scientists develop the materials and teach together.

The result? Students start to view ethics as a normal part of writing code rather than an afterthought.

“In a philosophy class, I'll identify a problem at a very abstract level, and people will ask, ‘What do we do about it?’ I'm like, ‘That's not my job.’ But co-teaching, the CS faculty will be like, ‘How do we make this better?’ Stakeholder analysis comes not from philosophers on high; it comes from very ethically thoughtful people that are in design situations, actually having to engineer things, actually worried about this stuff and generating workable procedures,” he said.

Now, embedded ethics is at the U, and in Fall of 2026, Nguyen is co-teaching this cross-listed class with associate professor in the Kahlert School of Computing Vivek Srikumar. In Spring of 2027, assistant professor of Philosophy Konstantin Genin is co-teaching with Jeff Philips. The class is expected to be very popular, with a mix of students from many different majors. So far, about half the students are in computer science, and half the students come from other majors such as philosophy, psychology or communications.

Students were already asking these questions

Nguyen said the questions students are asking in class have changed in recent years. They want to know if we’re losing our moral and social skills because of AI, whether there’s bias baked into the system or if there is systemic bias in AI hirings. “These are everyone’s favorite topics,” Nguyen said.

To illustrate particular ethical dilemmas in the AI space, Nguyen uses real-world examples, often starting the class with a case study about how machine learning was used to predict successful student outcomes in universities. “And the system is biased in really interesting ways, Nguyen said. “The system defined student success as graduation rate and graduation speed, but not in terms of skill or happiness or community. A lot of this is because the engineers only interviewed deans and above and did not talk to either faculty or students to get their views about what student success was.”

Another example is the famous COMPAS case, where a machine learning risk assessment tool used in the U.S. criminal justice system to predict the likelihood of recidivism was found to be racist. “The people building it weren't racist at all, and the code itself isn't racist,” Nguyen said. “But the world reflects past racism. And so that's going to be picked up in any database correlative system.”

Underlying these, and most AI ethics issues covered in the class, is the concept of value-ladenness, a philosophical term describing technologies and methods that appear to be neutral, but make very value-laden, politically charged decisions because of the design hidden underneath.

The class will also tackle the issue of moral and social deskilling: what happens when people rely on automation so much that they lose skills. The use of ChatGPT as a therapist illustrates the risks. “ChatGPT is not a licensed therapist,” Nguyen said. “ChatGPT,  as many people know, is measurably sycophantic. It tends to support people, and this is actually a huge difference between using ChatGPT as a therapist and a real therapist. A therapist will call you out; they will challenge you.” The worry here is that AI, as it’s currently designed, could cause people to lose the skill of hearing criticism.

“Criticism is what makes us improve,” Nguyen said. “Criticism is how we see perspectives outside of our own. As John Stuart Mill put it, criticism is how we know our thoughts are alive instead of dead dogma. One of the early projects that started pushing me to the social media space was an analysis of what echo chambers were, and the pernicious effect of having everyone agree [with] and praise your views. Making a machine that does that automatically? I am way more worried about that than killer AI.”

Inspiring goodness

To Nguyen, the headline news story of killer, world-ending AI is only one of 1,000 worries, and he reads it as maybe even a soft marketing campaign. “If you focus everyone's attention on the potential evil of superintelligent AI, no one's going to notice job displacement or the eroding of social life,” he said. “Somehow in the media, all the attention goes to the dangers of superintelligent AI, which, if you'll notice, is also really good for Silicon Valley, because behind that worry is the view that it's going to be super effective.”

There’s no denying that artificial intelligence is a powerful technology that will grow in ways that we can’t yet imagine, and the hope is that this class will train students who will eventually have a hand in its development.

Nguyen is hopeful about the ways his students could someday design this burgeoning technology for good; however, different people and countries often don’t agree on what ‘good’ means. Regulations and international agreements will likely help mitigate the risks of AI, but according to Nguyen, building ethics into the code is a more proactive approach. “Regulation can force people not to be terrible,” he said. “But a classroom can inspire people to be aggressively good and sensitive.”