Diverse data, ethical use key to responsible AI engineering
Karen Panetta's engineering journey led her to identify critical bias in facial recognition systems and champion transparent technology that protects human welfare.
The path for Karen Panetta's engineering degree started with her father's advice that her shopping habit needed to be sustained by a strong financial career.
After spending some time building computers and designing CPUs, the IEEE fellow and dean of graduate engineering education at Tufts University sought to go into academia. However, she realized that there was a big problem in the engineering field.
"They said, 'We're hiring you to be a mentor for women,' and then when I got to my first class, there were no women,'" Panetta recounted on the latest episode of the Targeting AI podcast from Informa TechTarget. "That's how I became an accidental mentor, proponent and champion to bring inclusive audiences to engineering."
That passion later transferred over to AI technology when she started to work to rescue children from human trafficking.
"I noticed all the data sets that were trained out there for facial recognition were taken straight on, both ears showing," Panetta said. "The database was mostly taken from people of color and criminal databases. So that was like, OK, there's bias right there."
In response, Panetta and her team created a diverse database of people from all over the world, ages 3 to 78. The images have people in different scenarios, either with sunglasses, masks or different head rotations.
"AI trained with very pristine data is not going to work in the real world," Panetta said. "Real world does not automatically translate to AI."
Panetta added that AI technology also needs to be ethical, which means not harming people or their livelihood.
"When we talk about ethical AI, we're talking about how we collect it, the rights of the people and how we're going to use it," she said.
Meanwhile, responsible AI is about being able to explain why the AI system made the decision it made, she said.
Esther Shittu is an Informa TechTarget news writer and podcast host covering AI software and systems. Shaun Sutner is senior news director for Informa TechTarget's information management team, driving coverage of AI, analytics and data management technologies, and big tech and federal regulation. He is a veteran journalist with more than 30 years of news experience. Together, they cover the Targeting AI podcast.