UMES granted $327K to study facial recognition bias
PRINCESS ANNE, Md. – Technology has improved many areas of life, but some argue it does more harm than good in other ways.
University of Maryland Eastern Shore students are set to study that issue with new grant funding. The research grant Facial Recognition and AI Tools in Criminal Justice – Building Accountability and Trust is led by Dr. Sandeep Gopalan, the Principal Investigator, and Dr. Chi-Fang Tsai, the Co-Principal Investigator. It was awarded by the Maryland Governor’s Office of Crime Control and Prevention.
These UMES researcher will use $327,000 to figure out what role facial and the impact recognition plays in the justice system. “Facial recognition technology is being used by the various law-enforcement agencies across the state,” said Dr. Sandeep Gopalan, with the University of Maryland Eastern Shore. The project also aims to build systemic knowledge about the use of facial recognition, bolster public trust, and establish guidelines for accountability.
The research team is studying the way that facial recognition works and just how often it could be wrong. “Being used in public gatherings, what if somebody is organizing a protest or in a shop or a shopping center because it can be used to prevent shoplifting but imagine if it misidentifies people who are innocently going shopping,” said Dr. Gopalan.
Dr. Sandeep Gopolan says through preliminary research, one demographic is more likely to get the short end of the stick. “Minority groups particularly people of color have been misidentified because of the biases in the algorithm,” said Dr. Gopalan.
Researchers at the HBCU say to bridge that gap, it can start with their students who are in the minority to funnel into the field of technology. By getting more people of color involved in developing tech like this, that diversity could help them create a fair system. “To address bias in our AI technology models and address the parameters and create a better model so that we can decrease the bias,” said Qian Leng.
For criminal investigations, this could mean the right person is on trial. “Analyze the data with a view to achieve some kind of system change for instance through law reform,” said Dr. Gopalan.
The research is set to be complete by this spring semester. Classes will be offered in the summer to analyze the data.
UMES is also developing a class for law enforcement and lawmakers to cover how life-changing this technology can be. Researchers also plan to prepare a report with possible reform recommendations.
