Point out that IBM is using photos on Flickr for face recognition technology without permission of the user



Face recognition technology is a technology that uses a camera and an algorithm to identify a human face, determines information such as gender, age, race, etc., and also identifies people who appear in photographs. Face recognition technology is still in the development stage, and companies are conducting various tests to improve their technical level. Under these circumstances, it turns out that IBM, a leader in face recognition technology, is using photos on the community site " Flickr " for the purpose of sharing photos for face recognition technology. It is pointed out that users do not agree to the use of photos for face recognition technology.

Facial recognition's 'dirty little secret': Millions of online photos scraped without consent
https://www.nbcnews.com/tech/internet/facial-recognition-s-dirty-little-secret-millions-online-photos-scraped-n981921

In January 2019, IBM released a data set called "Diversity in Faces" containing one million face data. This data set aims to improve the fairness and accuracy of face recognition technology and includes face images of different genders and races.

IBM releases huge data set containing 1 million face data aiming at "fairness of face recognition technology"-GIGAZINE



The DiF database is a collection of 100 million images tagged with " Creative Commons (CC) " on Flickr, and from the data set called " YFCC 100M ", images that meet the purpose of face recognition technology are further added. The thing that was carefully selected. Many of the people who tagged Creative Commons in the photo were to allow non-profit organizations and artists to use their own photos, but the posted images would be collected as a DiF dataset. There are also some who claim that "we did not assume that images would be used for face recognition technology."


By halfpoint

IBM claims that "the purpose of the research is to use photos," but that "IBM is not a charity, it is using data to gain in technology, so it may be an obvious commercial use." It is also being done. IBM sells technology to the New York City Police Department to search for surveillance camera footage from specific skin colors, hair colors and races. In addition, this surveillance camera image search technology is controversial because it is criticized for threatening civil liberties and for being against discrimination against certain ethnic groups.


by Daniel Naish

IBM responded that the company's privacy policy addresses the user's request to "delete images from the DiF database" as much as possible. However, when a man with more than 1000 photos on the DiF database submitted a request for deletion of his photo to IBM, there was a response that "user ID alone can not delete it." In order to delete an image from the DiF database, it is necessary to submit the URL of the image on the DiF database to the IBM side, and IBM has not published the list of photos included in the data set. After all, it seems that more than 1000 photos of this man remain in the DiF database. Also, the image is not deleted from the DiF data set and "YFCC 100M" data set that more than 250 organizations that have used DiF have so far, and deletion of the image is almost impossible. .

Not only IBM is collecting images from the Internet for face recognition technology, but many research institutes create data sets that collect images in the same way.

in Security, Posted by log1k_iy