The controversial surveillance company has now collected about 10 billion pictures from users of Facebook, Linked In, Instagram, and other social networks. The company uses these images to improve its algorithm which is used for facial recognition. Police and governments of various countries, including The Netherlands, have used the facial recognition software to search and identify potential suspects. By combining the collected images with progression in machine learning Clearview created a powerful, but dangerous tool. In an interview with Wired, Hoan Ton-That, one of the founders of the company, demonstrated the technology. With a smartphone app, Ton-That took a picture of the reporter. Within seconds, the app generated over 50 images from social network websites, showing the correct picture in each of them.
The software has sparked enormous outrage and fuels the debate of privacy in the era of social media and tracking online behavior. The American Civil Liberties Union has sued the company under a law that controls the collection of biometric data. Social Media companies Facebook and Twitter also raised their concerns and demand Clearview to stop their activities on the platforms of the social network companies.
The technology does not only scrape and compare billions of images from social media platforms, it also uses machine learning to deblur and predict the covered parts of a person’s face by filling in the missing parts based on patterns of the visible part of the face and other images. Experts say this increases the risk that a suspect is incorrectly identified and rises the chance of biases inherent to the technology.
References
https://www.wired.com/story/clearview-ai-new-tools-identify-you-photos/
