Designing Smart, Sustainable Cities Through Ethical AI Q&A with NRT trainee Connor Phillips, UT Austin

After finishing a master's in urban planning, Connor Phillips knew he wanted to pursue research that brought a more cutting-edge perspective and technology-based tools to the work of designing smarter, sustainable cities. When he discovered the work of Junfeng Jiao, NRT program director at The University of Texas at Austin, and learned about the newly launched NRT program on ethical AI, he knew he’d found his match. After just a year, Phillips has already published some of the first research papers on the practical application of generative AI in the work of city planning, exploring the potential benefits it could bring for residents, communities and society as a whole.

Conor Phillips, NRT Trainee at UT Austin


Talk about what brought you to UT Austin and to NRT.

I first learned about Junfeng Jiao when I was working on my master’s thesis at the University of California, Irvine. I read some of his work on AI, exploring how people view cyberspace as being interrelated to physical space and thought it was fascinating. At the time, I was working on a lot of urban mobility tools like scooters, but I was beginning to feel drawn to other topics and directions like examining the intersection between digital and physical space in urban planning. Jiao’s work was really aligned with that interest. And when I had the chance to meet him, he explained that the NRT program he was launching with a number of other professors focused on the ethical use of AI. It was just getting started, but I’m absolutely thrilled that my timing fit with the launch. It’s really been a wonderful opportunity so far to pursue my research with strong mentorship.

What research areas are you focusing on right now?

I’m working on a number of projects all related to generative AI, things like ChatGPT (a large language model-based chatbot) and DALL-E (an AI system that can create realistic images and art from a description in natural language) that will produce written and graphic content. I don’t come from a computer science background, so the work I’m doing is really focused on accessing AI technology in a more straightforward way and applying it in practical ways to urban planning work. For example, I started with a project that involved feeding language about redesigning a neighborhood into DALL-E using buzzwords that come up in those kinds of projects — like mixed-use development — and then asking the generative AI technology to produce images of the proposed neighborhood.

Those kinds of images can be incredibly compelling when pitching a project to community stakeholders, but they can also be incredibly expensive to produce, especially when you have to hire a design firm. That often creates a real barrier for those who are trying to develop projects that may not bring in a lot of income, like a low-income housing project. AI tech can help to eliminate that barrier. It’s not perfect by any stretch, but it's a way to increase equity in the way planning projects are considered in different communities.

How has being a part of NRT supported your research?

I think having access to a range of professors outside of our department is hugely important. The PIs and co-PIs in this program are spread across departments, and that’s given me the chance to tap into so many different perspectives. I’ve also been able to build the knowledge I need in computer science to pursue research in the cross-section between urban planning and AI. That’s been critical for me. UT’s urban planning program is more traditional in terms of its research areas. It has foci on housing and transit, whereas my interests lie in smart cities. The NRT and Urban Information Lab provide me with additional important resources to pursue this research.

Part of NRT focuses on professional development and helping you prepare for future career paths. How has that been a part of your experience so far?

Jiao’s mentorship is a big part of it. He’s definitely focused on helping to launch my career by supporting me in putting out as many papers as possible on the interconnection between AI and urban planning. There isn’t a lot of scholarly work in this space yet. Usually when you search for something on Google Scholar, there’s at least 10,000 entries, but if you enter ChatGPT and urban planning, there are just a few hits. Because I want to go into academia, the reality of publish or perish is very real, so it’s incredibly valuable for me to be crafting some of the earliest papers on this topic.

At the same time, I really appreciate NRTs focus on providing professional development opportunities outside the academy, and I certainly want to take advantage of that. I’m interested in pursuing an internship at a think tank-type organization to get a different perspective and viewpoint on this work. I think bringing that back to my academic work will better prepare me for research and teaching in the future.

What might you say to a prospective trainee who is thinking about pursuing NRT?

I’d encourage them wholeheartedly. I’ve really found no downsides to this experience! I’m actually more attached to my NRT community than I am to any particular school or department at this point. I feel like I'm part of a group of folks who are really committed to research that pushes boundaries quickly. Urban planning is not a field that tends to move very fast classically. Being a part of NRT allows us to be more on the cutting edge. Not only am I able to produce work at a much faster rate, but I’m also able to explore work in more varied and interesting directions. When it comes to AI, it’s crucial that we take that approach to research because it’s expanding and developing so quickly.

And there’s so much intentional collaboration as well. Jiao and his team are working to identify common research themes for those of us in his lab and encouraging us to start working together on research projects. Pursuing a doctorate can be a really solitary endeavor, so it’s been inspiring to get drawn into a supportive and collaborative community. The support of such a strong and interdisciplinary team allows our research to move quicker than I ever thought was possible.