Teaching Responsible Approaches to Tomorrow’s AI

Brinnae Bent joined Duke Engineering this past summer as an executive-in-residence in its Masters of Engineering in Artificial Intelligence program. In this role, she teaches several courses in both the residential and online programs and serves as an advisor to students. Bent is also the instructor of Duke’s newly launched Coursera Specialization online course on Explainable Artificial Intelligence (XAI).  

Originally from Flint, Michigan, Bent moved to North Carolina to pursue her undergraduate degree in engineering from North Carolina State University. However, her journey to AI has been anything but linear.

In high school, as a nurse assistant in a nursing home, Bent witnessed a wide range of problems within the healthcare system. She saw issues like wheelchairs that would not work and alarms that would fall off. These experiences at a young age ultimately led Bent to study biomedical engineering at NC State to provide solutions to these real-world problems.

After falling in love with research during her undergraduate degree, she pursued a masters in neural engineering at Duke. She then decided to stay at Duke to pursue a PhD in applied machine learning to continue her work in solving pressing issues in healthcare.

“I have always seen AI as a way to help solve big problems – from healthcare diagnostics to making new scientific discoveries to helping improve sustainability,” Bent said.

However, her passion for AI does not stop at healthcare. Bent collaborates on a variety of research efforts around Duke’s campus and mentors students on AI alignment research. Outside of Duke, she is a trustworthy AI consultant, where she works with companies on the responsible development and application of AI technology.

At its core, her work and interests revolve around the creation of ethical AI and its importance. She said it is crucial to develop technologies like AI that are responsible and whose outcomes align with human goals, as the opposite is detrimental.

“If engineers don’t know how to robustly evaluate a machine learning model, they may ship it to production and cause problems for real people,” she said. “If they don’t think about the ethical consequences of their algorithm development while in the classroom, they won’t when they are out building ChatGPT 2.0, either.”

She said one of the reasons she decided to work at Duke was its focus on developing AI responsibly.

“I wanted to be a part of a community of people who are strong believers in ensuring the technology is developed ethically, and I have found that community here at Duke,” she said.

Bent also said that Duke has been a “force” in AI for a long time and that she particularly “loves the multidisciplinary feel of machine learning at Duke.” Duke has researched and applied machine learning all across the community, from healthcare to philosophy to media arts.

“AI is much more than ChatGPT – this technology can help humans solve some of the most critical problems of our time,“ she said.

When asked about her future plans, Bent has a unique outlook.

“I stopped making plans for the future a while ago,” she said. “Instead, I try to spend my time exploring what I am curious about. I think that makes for a much fuller life.”

Outside of her work, Bent is a painter and uses her artistic expression to showcase the interactions between humans and AI. She is also an “ultrarunner,” in which she recently ran 50 miles in the Umstead 50 Race. While she is currently on a break from running, Bent said she does plenty when chasing her two-year-old son around.