Facebook AI is releasing 100,000+ videos as a training set for its $10 million challenge to help detect deep fakes. Each video presents real people with a fake counterpart.

The challenge is to create automated tools that can identify deceitful videos and manipulated media.

Facebook created the training set by soliciting videos from everyday people doing everyday things Participation is open worldwide through March 2020.

Examples from the Facebook training set for deep fake videos

Other organizations working on training sets or automated tools for deep fake detection include DARPA, the AI Foundation, and Google.

The Facebook AI Detection Challenge was first announced in October, 2019. Presenting participants include Amazon Web Services (AWS), Microsoft, and the Partnership on AI. Also involved are academics from Cornell Tech, MIT, University of Oxford, UC Berkeley, University of Maryland, College Park, and State University of New York at Albany.


Facebook AI Launches Its Deepfake Detection Challenge
IEEE SPECTRUM | December 11, 2019 | By Eliza Strickland