Photo: Wired.
That's true, but with the "10 years ago" image movement, the user's previous and current photos posted with a specific timeline and exactly 10 years. And users often post in public mode.
Imagine, in order to be able to create a face recognition algorithm for age-related traits, in particular the aging process, you will need huge and detailed data warehouses. And what's better if you have a database of people changes in a specific time to train machines.
Of course, it is possible to exploit the images that users have posted by viewing the posted date or EXIF data. However, these data may be inaccurate because images may be posted incorrectly with shooting time. EXIF data may also be invalid because some images are downloaded from other platforms, screen capture .
In addition, not all users use personal photos to represent and not all images on the timeline are related to the face, they can post a friend, cartoon character, animal image. .
In short, the trend of "10 years ago", users are willing to share images with the exact time of shooting, accompanied by a detailed description of the time, place, person . This helps create a data warehouse. Image data is clean, simple to use. In addition, they have a hashtag # 10yearschallenge, which makes data collection easier.
Sophisticated facial recognition algorithms will remove non-face images.
Kate O'Neill said that this is not the trend, the only game taking advantage of users to collect data. The most specific example is the Cambridge Analytica data collection of 70 million Facebook users in the US.
The data is used for many purposes
So who is behind the collection of photos of Facebook users to train face recognition algorithm?
The answer is no, because this is something that users must accept when using social networks. But what is worth mentioning here is how these data are used.
Kate O'Neill gave three theories for using these data.
1. Good scenario
The ability to identify changes by age can be helpful in finding missing children for long periods of time, when their appearance has changed from the beginning.
Last year, thanks to facial recognition technology, police in New Delhi, India found 3,000 missing children in just 4 days.
2. The script is not good
Targeted ads will take advantage of face recognition technology to analyze your aging speed. Since then, insurance companies and health care companies can rely on data collected to refuse to sell or pay more because your aging rate is too fast.
User monitoring script
In 2016, Amazon launched a real-time face recognition service and sold it to the government, law enforcement agencies like police departments in Orlando and Washington County, Oregon.
Police can use this technology to monitor anyone they consider annoying, including those who do not commit crimes. Because of this, the US Civil Liberties Alliance has asked Amazon to stop selling its real-time face recognition service.
Regardless of the origin and purpose of the "10 years ago" trend, users need to know more about personal data and how to give it to technology companies.
Currently, users and their personal data are becoming a big, "lucrative" data source for most technology companies. In order to protect and avoid possible unfortunate risks, users need to treat their data well.
See more: