You might remember FaceApp — a selfie-editing app that transforms users’ pictures by making them look older, younger, or giving them an artificial smile. Well, earlier this month, the Russian app’s creators updated their software with four new filters: “Asian,” “Black,” “Caucasian,” and “Indian.” First spotted by Mic, users can select these options and the app will alter their appearance, changing the color of their skin and hair.
It’s tantamount to a sort of digital blackface, “dressing up” as different ethnicities. Snapchat was previously criticized for implementing a similar feature when it released a Bob Marley selfie mask to celebrate 4/20. FaceApp itself has been criticized for racial insensitivity in the past, with users pointing out that its “hot” filter consistently lightened users’ skin tones.
The company later apologized for the feature, with CEO Yaroslav Goncharov explaining that the effect was an “an unfortunate side-effect of the underlying neural network caused by the training set bias.” This means that the software that was used to change users’ appearance had been fed only pictures of white people, so this was the skin color it associated with “hotness.” This sort of data-led bias is a big problem in the field of artificial intelligence, with programs regularly embodying racial and gender prejudices because of the data they’re trained on.
In the case of FaceApp’s latest update, though, Goncharov claims there is no bias or prejudice involved. “The ethnicity change filters have been designed to be equal in all aspects,” he told The Verge over email. “They don’t have any positive or negative connotations associated with them. They are even represented by the same icon. In addition to that, the list of those filters is shuffled for every photo, so each user sees them in a different order.”
Some might point out, though, that the order the filters appear in is hardly the issue — it’s the fact they’re being offered in the first place.Read more at theverge.com