Penn State profs warn of 'racial bias' in Google Arts app
- Google's popular new feature allowing users to match their faces to famous portraits has raised privacy concerns due to its facial recognition feature, but some Penn State professors are more worried about the app's possible "racial bias."
- The professors contend that the app does not have enough non-white faces in its database of portraits, but concede that this may be due to bias on the part of museum curators, rather than app developers.
While others express privacy concerns about the face-matching feature of the Google Arts and Culture App, some Penn State professors are fretting that the app “might have a racial bias.”
A new facial-recognition feature on the app that matches users’ pictures with the most similar classical portrait contained in the app’s database has already exploded in popularity, with several celebrities posting pictures of their matches, such as Kristen Bell, who was given a 32 percent match with John Pettie’s “August Manns.”
Some, however, think the app “might have a racial bias,” including Pennsylvania State University Professor Nancy Locke, according to The Daily Collegian.
“I’m seeing some concern among friends and people in the art world that the app might have a racial bias,” she remarked. “I think it probably just doesn’t have enough portraits in the database of non-white persons and that might be skewing the results.”
Instructor Farima Fooladi expressed similar concerns, noting that she and her Middle Eastern friends believe they weren’t accurately matched with portraits, though she didn’t necessarily see this as a bad thing.
“The app brings up a conversation about diversity to the front in the art community,” she elaborated.
Carolyn Lucarelli, visual resources curator in the Department of Art History, attributed the apps alleged lack of diversity to Google’s limited database.
“It’s going to take time. I know they’re adding all the time,” she commented, noting that since people of color are often underrepresented in museum collections, “it only makes sense that that’s what’s being represented in the app.”
Follow the author of this article on Twitter: @AGockowski