Sun. Feb 23rd, 2020

Facebook donated $10 billion to challenge deepfakes

2 min read

Under the blessing of open source projects, it is very mature to change the face of pictures and even videos. The face of artificial intelligence can be adjusted to the point of realism. Although this kind of face-changing is mainly concentrated in the technical circle field and is keen to deal with porn videos, it is obvious that technology brings fun and has the opposite effect. The most typical negative advantage is that it can be used to make fake pictures or video content, and then used to publish false news guides and even manipulate public opinion. Most ordinary users can’t judge whether the picture in the message is true or not. If it is a video, it will be more difficult to judge and the potential credibility is higher.

deep neural music network MuseNet

Facebook has announced that it has cooperated with many well-known institutions, and everyone will jointly train the model and then detect the false content of artificial intelligence. The use of fake content to detect is actually both what Facebook should do, therefore, whether it is from the user’s point of view or from the aspect of compliance, Facebook needs to actively combat false content, and the content of artificial intelligence is the focus. Facebook hopes that the industry can use innovative methods to detect and prevent media manipulated by artificial intelligence to mislead or use others, guide or manipulate public opinion.

It’s still very easy to make fake video or images with open source projects, but it’s very difficult to accurately detect these fake content. This is why Facebook is looking for cooperation with a number of well-known research institutions. For this project, Facebook will invest $10 billion to motivate institutional researchers.

Companies or organizations involved in the project are also among the best in the industry, such as Microsoft Corporation. Cornell Tech, MIT, University of Oxford, UC Berkeley, University of Maryland, College Park, and University at Albany-SUNY. will also participate. An official Facebook spokesman said that he hopes to promote more research and development in the industry through this project, ensuring that open source tools can detect all kinds of false content.

Via: Reuters