From CBS SF Bay Area
To counter prevalent concerns regarding the spread of deepfakes, or videos doctored with artificial intelligence, Facebook is offering a unique solution: make even more deepfakes.
The social media company is paying for the creation of its own deepfake videos to create a data set. It believes that the data set, which will include source videos and imagery of actors, will assist experts in the artificial-intelligence community in their pursuit to develop methods to expose deepfakes.
Facebook is sponsoring the competition, Deepfake Detection Challenge, and offering grants and awards to encourage participation from AI researchers. The company is collaborating with various organizations, including Microsoft and the University of California, Berkeley, and putting up more than $10 million.
Hany Farid, a professor at UC Berkeley and image-forensics expert whose lab received a grant from Facebook related to its deepfake detection research, said the competition is a big step toward solving an important problem. He said Facebook will also need to keep in mind that any technological solution must change over time, similar to the ways technology advances for stopping spam and computer viruses.
“It’s always evolving because our adversaries are always evolving,” Farid said.
Hany Farid is a professor at the UC Berkeley School of Information and EECS. He specializes in digital forensics.