Provost column: What is liberal arts engineering?

If you have heard me speak about engineering and artificial intelligence at Carolina, you’ve probably heard me mention my excitement for liberal arts engineering and why I believe that we are perfectly positioned to launch and lead in this important, emerging field. But what does liberal arts engineering actually mean?

In a nutshell, instead of allowing AI — and technology more broadly — to be something that simply happens to us, liberal arts engineering puts humans in the driver’s seat. By grounding students in the liberal arts and humanities — focusing on and nurturing curiosity, creativity, storytelling and problem-solving — we can harness AI and other technologies to serve humanity through our professions, rather than allowing technology to dictate how we live and work.

To understand why we need to take the impact of AI seriously, it is worth noting how we got here. I happen to be a roboticist and have ridden the AI rollercoaster for longer than most. Here’s my abbreviated take on history. Rewind several decades, scientists were developing computational models inspired by the human brain, resulting in a structure that became known as the artificial neural network. Researchers were able to say things theoretically about how these networks might behave, given enough training. But back then, there simply was not enough data available to train them properly. Nor did we have the computing firepower needed even if the data had been there.

Fast forward a few decades and the raw computing power had expanded significantly. The internet provided an almost unlimited amount of training data. This confluence of data and compute allowed us to build and train networks with enough layers to be useful. They became known as deep networks, and all sorts of deep machine learning emerged, like deep reinforcement learning, deep recurrent neural networks and so on.

The next breakthrough happened just a few years ago, and it had to do with a slight tweak to the architecture itself. Rather than having similar types of nodes everywhere, it proved useful to layer the network with different kinds of nodes. People developed architectures that better captured the relationships among words, images and other data across long sequences. These so-called transformer-based models proved key to producing large language models and generative AI agents, which is more or less how we got here.

But the story about AI is only partially a story about technology. It is also an adoption story. ChaptGPT gained tens of millions of users in a matter of days, making it one of the fastest-adopted consumer technologies in history. In fact, many economists and technology scholars already consider AI a general purpose technology, like the steam engine or electricity or the internet. What this means is that AI is not constrained to any one sector of society or to any one business segment. Its impact is vast and hard to predict, already changing the way we interact with each other and many facets of society. Take electricity for example: It didn’t stop at light bulbs; it changed how we live. Skyscrapers could not exist without elevators. And elevators could not exist without electricity. Our built environment is different because of electricity. Which raises more questions: What will the cities of the future look like because of AI? What comes after the skyscraper?

No one really knows the answers, nor what precise skills will be needed to be successful in the future AI economy. Given all this uncertainty, how are we approaching all of this at Carolina? How do we prepare students for a world that literally does not exist yet?

I believe that what we need are people who can see around corners. Who can make new connections. Who can go from information to wisdom to judgment. And who can keep a finger on the pulse of the future with limited and oftentimes noisy and even contradictory information. And where in the curriculum does this live? Where do you see around corners? This is what the liberal arts are all about.

In this context, we talk both about “learning to read” and “reading to learn.” The learning to read piece is about AI literacy, and we have now structured our curriculum so that starting this fall, all of our incoming students are exposed to multiple iterations of AI literacy modules and content early on in their education. The reading to learn, or AI competency piece, is the domain- or field-specific use of AI, which we have all throughout our curriculum, like AI in healthcare, AI in computer science, AI in law, and so on.

In support of this, we recently announced that professor Saif Khairat from the School of Nursing has stepped into the role of vice provost for AI here at Carolina. I am excited about working with Saif in this role; a University-wide position is much needed as AI is already intersecting with almost all aspects of our enterprise, from how we teach and conduct research to how we run our internal operations.

At Carolina, we foresee that the liberal arts are about to see a major renaissance — not as a standalone but as a necessary and fully integrated companion piece to the more technical aspects of AI competency. It is my ambition as provost to see the new discipline of liberal arts engineering be born and take off on Carolina’s campus over the next few years. Stay tuned.

Provost’s log, stardate July 13, 2026