In what could be termed as a major infringement to online privacy, IBM has taken nearly a million photographs from photo-hosting site Flickr, without the knowledge of the users and used them in their facial recognition training programs, the NBC News reported.
IBM also shared the collection to researchers as a key step toward reducing bias in facial recognition.
Many photographers whose photographs were included in IBM’s data research were not informed that their photos had been used and detailed with facial geometry and skin tone and they may get used to develop facial recognition algorithms.
“None of the people I photographed had any idea their images were being used in this way,” the report quoted Greg Peverill-Conti, a Boston-based public relations executive who has over 700 photographs in IBM’s collection, known as a ‘training dataset’
In a statement to The Verge IBM has said, “We take the privacy of individuals very seriously and have taken great care to comply with privacy principles,”
The statement said that the dataset could only be accessed by verified researchers and only included images that were publicly available. It added that, “Individuals can opt-out of this dataset,”
IBM is among several companies which are exploring the field of facial recognition and is not alone in using photos of individuals without asking for their consent. For instance, Facebook has photos of 800,000 faces open for other researchers to download, the report said.