IBM faces privacy row for using Flickr in face-recognition project
13 March 2019 14:11 GMT

A report by NBC News has revelaed that IBM has been using around one million images from the image-hosting platform Flickr to train its facial recognition AI, without the permission of the people in the photos.

In January, IBM revealed its new "Diversity in Faces" dataset with the goal to make facial recognition systems fairer and better at identifying a diverse range of faces — AI algorithms have had difficulty in the past recognising women and people of colour.

IBM said in a statement that it had taken great care to comply with privacy principles.

But one digital rights group said IBM's actions represented a "huge threat" to people's privacy.

"None of the people I photographed had any idea their images were being used in this way," a photographer told NBC News.

Photos selected by IBM were listed under a Creative Commons licence, which generally means the images can be widely used with only a small number of restrictions.

In a paper published online about the work, IBM researchers describe in detail the steps taken to analyse people's faces, including taking measurements of the distance between individuals' facial features.