We all lament the general loss of Online privacy but a new & extremely detailed study of facial-recognition data shows how much the rise of deep learning has fueled the loss of privacy.
The study by Deborah Raji, a fellow at non-profit Mozilla, & Genevieve Fried, who advises members of the US Congress on algorithmic accountability shows just how much the business of facial recognition has eroded our privacy. In one sense the research shows up the ill effects of facial recognition.
A report in the MIT Technology Review by Karen Hao, while refrring to this new study said this hasn’t just fuelled an increasingly powerful tool of surveillance but the latest generation of deep-learning-based facial recognition had “completely disrupted our norms of consent.”
During their study the team examined over 130 facial-recognition data sets compiled over 43 years. They found that researchers, driven by the exploding data requirements of deep learning, gradually abandoned asking for people’s consent. This had led more & more of people’s personal photos to be incorporated into systems of surveillance without their permission even.
In 2007, the release of the “Labeled Faces in the Wild” (LFW) data, said Karen, opened the floodgates to data collection through Web search. Researchers began downloading images directly from Google, Flickr, & Yahoo without concern for consent. LFW also relaxed standards around the inclusion of minors, using photos found with search terms like “baby,” “juvenile,” & “teen” to increase diversity.
According to the MIT write-up, Raji said her investigation into the data had made her “gravely concerned” about deep-learning-based facial recognition. She now hopes the paper will provoke researchers to reflect on the trade-off between the performance gains derived from deep learning & the loss of consent, meticulous data verification, and thorough documentation. Was the need to do deep learning worth abandoning all of these practices, she wonders?
