Clearview AI built its facial recognition system by exploiting your “publicly available” social media profile photos.
They were able to do this because most social media companies make profile photos public by default and deprive users of the option to hide them. Thread:
They were able to do this because most social media companies make profile photos public by default and deprive users of the option to hide them. Thread:
On Facebook, "your current profile picture and cover photo are public, so they can be seen by anyone on or off Facebook." There's no setting to change this. Facebook won’t let you protect yourself. facebook.com/help/193629617…
On Instagram, “there's no way to hide your bio or profile image." help.instagram.com/347751748650214 Even if you have a private account. The company adds: “if you don't want people to see this info, we'd suggest removing it from your profile." Sorry not sorry.
On Twitter, "the profile information you provide ... is always public, like your biography, location, website, and picture." Got a private account? Your profile photo is still public. So, tough luck. help.twitter.com/en/safety-and-…
LinkedIn makes profile photos public by default, too. linkedin.com/help/linkedin/…
But unlike the other major networks, LinkedIn allows you to make it visible only to connections: linkedin.com/public-profile…
But unlike the other major networks, LinkedIn allows you to make it visible only to connections: linkedin.com/public-profile…
Key social media companies - through default settings & a lack of profile photo privacy controls - allow user face photos to become raw material of surveillance systems.
Users should not have to avoid using a photo of their face to keep themselves out of a dystopian face recognition database. And the reality is that many people, such as workers or job seekers on LinkedIn, can’t avoid using a photo of themselves. This is not users' fault.
Our team at the ACLU (cc @nicoleozer @snowjake) has repeatedly highlighted the vulnerability of profile photos. Social networks knew this was a problem years ago: nytimes.com/2018/10/12/tec…
Clearview isn’t the first and it won’t be the last company of its kind. The social networks should publicly explain what they will do. Not by using horrible laws like the CFAA, but by adding privacy controls & design that will actually protect users.
