Faculty Spotlight: Dr. Jiadong Lou

New faculty member Dr. Lou is in search of our secure future as artificial intelligence becomes more and more a part of our lives.

Lou’s research focuses on the analysis of security risks and data privacy in machine learning, aiming to understand and mitigate how AI systems may expose or misuse user information as they become deeply integrated into daily life. "You use ChatGPT everyday, and then you are typing sensitive information to let them give you the answer. And that will become a big problem, whether such kinds of generative AI or other kinds of models will collect your data for training purposes without your consent,” he explains.

In the past couple years, companies including The New York Times, Disney, and Sony Music Entertainment have filed lawsuits against AI firms for using their content illegally to train their generative models. In his research, Lou is looking for the kinds of auditing and detection methods that can be used to find evidence that can be used in court, for example, to determine whether a company’s dataset or model has misused private or copyrighted information during training.

His solution? To map out how the “black box” truly operates. “We need to provide rigorous mathematical evidence to quantify how much a model may rely on or use our data,” Lou explains. Rowan’s collaborative environment has given him the opportunity to work with professors in the mathematics department on this effort. “We have many excellent researchers in optimization and statistics who help us analyze the internal behavior of machine learning models, rather than relying solely on surface-level observations,” he adds.

Due to the matter’s gravity, the research walks a fine line. “It’s challenging because we must protect both users and the companies providing AI services. On one hand, we need to detect potential misuse of data; on the other, we must also ensure that innocent models are not wrongly accused. We cannot make rough or biased judgments about whether a system has used private data.” His extensive works, which covers similar topics in machine learning security and privacy, can be found here.

In previous years, Dr. Lou’s research has spanned several domains, including network optimization, mobile application analysis, and the use of AI in geoscience. With many of his diverse works published in top-tier conferences and journals, Dr. Lou emphasizes that the true goal of research is not publication itself, but producing work of lasting value. “We may not publish every paper in a top conference, but we should always aim for that level of rigor,” he says. “The goal is not simply to get a paper accepted, but to conduct research that truly matters. I tell my students to dream big, and then do something big.”

This spring, Dr. Lou will be teaching a one-of-a-kind special topics course at Rowan, incorporating elements of machine learning with elements of cybersecurity. This class will be offered in both an undergraduate and graduate format.

Any students looking to learn more about him or research under him can find his information at his faculty profile or his personal website.