Over last couple of months major internet companies including Facebook and Google have come under pressure regarding their Face Recognition technology. While the technology was initially used for tagging family members and friends on different images, Facebook are using the technology for other purposes. Interestingly, Facebook’s researchers are now working on developing AI to combat other AI that might be trying to identify and tag people in videos.
This news comes as a lot of internet users have expressed their concerns regarding the face recognition technology. While the technology can be pretty useful, the recent increase in deepfake apps and services has shown that the technology can be used to spread false information while compromising users’ privacy and security.
While there are AI that can prevent this, the technology is limited to images and interestingly, researchers from Facebook have developed a new system that will be able to detect this for videos as well as live streams.
In case you are interested, the AI uses an adversarial autoencoder to alter someone’s moving face in a video. Although we won’t be able to spot the subtle differences, face recognition software will be able to identify the problems and will be able to identify any kind of alterations.
At the moment however, it’s unclear how or where this technology will come handy. Despite that, as a large number of companies are implementing face recognition technologies, it will be interesting to see how the latest research develops in the coming years.