Facebook and Microsoft join forces to identify deepfake videos

The social media giant will invest $10m in “Deepfake Detection Challenge”. As part of the Facebook project, it will allow researchers to produce real denfaques and create data sets for testing identification tools.

The company said the videos will be launched in December, part of which will be paid actors and that no data of users will be used.

As the American presidential election approaches, social media platforms have been put under pressure to address deepfaque concern that uses artificial intelligence to produce real video where a person appears to say or do something that is not true.

Technology is also becoming more accessible to less skilled creators. Last week a Chinese app called Zao made it possible for users to place their faces on some video clips.

Deepfake Detection Challenge is not the first time that Facebook, which does not apply a particular policy for deepface videos, finances academic research for it.