Uncovering bias in artificial intelligence

Robots on books getty image

Photo Credit: Getty Images

By Kennedy Ryan
December 15, 2023

Peizhao Li, PhD’24, a candidate in computer science, has received a $55,500 Graduate Research Fellowship from the National Institute of Justice. 

He was selected as one of 24 recipients for his proposal, 'Regulating Data Bias in Intelligent Decision Making.’ The funding will support his research in artificial intelligence fairness – rectifying biased and unfair AI technology to make it non-discriminatory.

While artificial intelligence is used to make automated decisions and complete tasks for major corporations, many of the datasets used to build the software contain preconceived biases. These inputs can directly affect minorities. 

Artificial intelligence is actively being used by major corporations for predictions of all kinds, such as credit risk applications, fraud detection, and recruitment systems. These systems each impact decisions based on gender, race, and other forms of identification. 

“I think it’s very important to the United States as well as society,” said Li. “We talk about discrimination every day, but there are still many biases in technological innovations toward individuals. We need to work towards a more fair society.”

Throughout his PhD program, Li has worked alongside his advisor, assistant professor Hongfu Liu, to examine bias within artificial intelligence datasets. They determined that when developing these artificial intelligence systems, cleaning datasets and looking for these biases in advance of processing information can reduce these incidents. 

“I think AI can be used more responsibly. It’s important for us to raise concerns at the earliest stages,” said Li. “With this development we can expect a more safe and unbiased result.”