AWS Throws Weight Behind Deepfake Detection Challenge


Amazon Web Services (AWS) will work with Facebook, Microsoft and the Partnership on AI, on the first-ever Deepfakes Detection Challenge, a competition designed to produce technology that can detect when artificial intelligence has been used to alter a video, in order to mislead viewers.

As part of the challenge, AWS will provide up to $1 million in AWS credits to researchers and academics over the next two years, along with competition data. AWS will also offer up the support of Amazon machine learning experts to help contest teams get started.

“The same technology which has given us delightfully realistic animation effects in movies and video games, has also been used by bad actors to blur the distinction between reality and fiction,” wrote Michelle Lee, VP of AWS’s Machine Learning Solutions Lab, in a blog post.

“’Deepfake’ videos manipulate audio and video using artificial intelligence to make it appear as though someone did or said something they didn’t. These techniques can be packaged up in to something as simple as a cell phone app, and are already being used to deliberately mislead audiences by spreading fake viral videos through social media.

“The fear is that deepfakes may become so realistic that they will be used to the detriment of reputations, to sway popular opinion, and could in time make any piece of information suspicious.”

The Deepfakes Detection Challenge is looking for different approaches to detecting deepfake audio, video and other tampered media, and will begin in in December with the release of a new dataset generated by Facebook, one comprising tens of thousands of example videos, real and fake. Competitors will then use the dataset to design algorithms that can detect a real or fake video. Those algorithms will be evaluated against a secret test dataset, one that will not be made available to contestants.

“Building deepfake detectors will require novel algorithms which can process this vast library of data (more than 4 petabytes),” Lee wrote. “AWS will work with DFDC partners to explore options for hosting the data set, including the use of Amazon S3, and we will make $1 million in AWS credits available to develop and test these sophisticated new algorithms.” Participants will be able to request a minimum of $1,000 in AWS credits to start, with additional awards granted in quantities of up to $10,000, as entries demonstrate r success in detecting deepfakes.

Participants can click here to learn more and request credits.

In addition to being a founding member of the Partnership on AI, AWS is also joining the non-profit’s Steering Committee on AI and Media Integrity.

The Deepfakes Detection Challenge steering committee will share the first 5,000 videos of the dataset with researchers working in the field.