Named as one of Pittsburgh’s 40 under 40, Haiyi Zhu is the Daniel P. Siewiorek Associate Professor of HCI and director of the HCI undergraduate program at CMU. Her research focuses primarily on the integration of AI systems in various high-stake environments such as child welfare and special education. Check out Haiyi’s class on Human-AI Interaction (05618/05318) that happens every semester!
Haiyi received her undergraduate degree in computer science at Tsinghua University and completed her masters and PhD program in HCI here at Carnegie Mellon University. Yet, prior to her degrees abroad, she was never exposed to HCI as a potential field as that was never available as a relevant course or program.
“When I tried to apply abroad for PhD programs, I discovered HCI and CMU. Even though I chose [to study] CS—actually, before that I had a math and physics degree—I found that those are too abstract, and I wanted to solve some real problems. But even in CS, I felt like I didn’t care about the computer [as much], I also cared about the human [aspect]. Then I discovered that HCI is such a nice interdisciplinary area that studies both computers and humans, so I was really excited.”
Haiyi discussed how she cared deeply about solving problems that extend beyond computation and algorithms. She wanted to understand and impact the human aspect of technology as well. That is how she landed at Carnegie Mellon to do her PhD within HCII.
From there, Haiyi rose within the department to her role today. She has quite a long list of responsibilities but estimates that she spends 30% of her time teaching, 30% managing the undergraduate HCI experience, and 40% of her time in research.
When she isn’t working, Haiyi actually loves watching zombie movies and TV shows. To our surprise, she is able to watch them right before bed and has no trouble falling asleep. We know whose office to head to if there is ever a zombie apocalypse!
Haiyi’s simple answer to this question is that she truly is a human-computer interaction researcher. Diving deeper, though, she studies human-AI interaction. She is focused on how we can integrate AI into society, making sure we are maximizing the benefits it can bring while mitigating any potential harm.
*“I have many lines of research, [but] I’m interested in AI within high-stake domains [such as] child welfare, human services, mental health and special education classrooms. How can we engage impacted community members in the design and evaluation process of these AI tools, and ***how can we work with people in power to make these tools actually work well and appropriately in the [desired] context?”
These are the high level research questions that Haiyi seeks to answer through her research. There is a huge need for artificial intelligence in these spaces to help get the right resources and care to the people who need it most. However, there are drastic consequences if AI approaches the problem incorrectly. Haiyi works to make sure that AI is being used in the right way to solve the right problems in the spaces that need the most help.
Haiyi notes that the answer to this question is very context specific—how the tools were designed, how they are used by their human counterparts, and how one should go about evaluating its performance are all important factors to consider.
“It’s very hard to say broadly what we should do at this point or generally how [AI] is affecting our society, so I always say that we should [first] talk about what this particular AI is doing in a particular context in this domain. For example, if we are talking about AI support for programmers, we should look at the impact on the coders, the impact on the companies and then see how we could design it.“
Investigating the unique environment in which an AI system is situated in is something that Haiyi often does in her research studies on Human-AI Interaction.