The DeFake project is an initiative by a team of researchers at the Global Cybersecurity Institute at RIT and School of Journalism and Mass Communications at UofSC.
Deepfake technology has slashed the amount of knowledge and time it takes for bad actors to generate believable malicious video content. These videos could lead to disinformation campaigns attacking democracies and ruin reputations of high profile individuals. The faith of the public in the news media could erode if these videos are published. The DeFake project combines user studies with journalists as well as the latest in the AI research to develop a tool that would help the journalists detect these videos. The team are one of eight winners of the Knight Foundation’s AI and the News Open Challenge.
The Team
Tool Limitations
The project is currently under constant development. The current limitations of the detection include:
- Only pieces of videos with the subject facing the camera can yield results
- The models lean heavier towards detecting videos as fake