DeepFake Videos a Threat to 2018 US Midterm Elections

Siwei Lyu, director of the Computer Vision and Machine Learning Lab at the State University of New York in Albany, recently wrote that fake videos made with machine learning called “deepfakes” may pose a significant problem to the upcoming 2018 US midterms.

However, the SUNY associate professor of computer science, along with colleague Ming-Ching Chang and Ph.D. student Yuezun Li said that they have found a reliable way to tell what is a fake video. The experts pointed out that the algorithm used for making the videos is still not perfect, which makes the simulated face unable to blink in a way that a real person does.

Nevertheless, Siwei pointed out that the technology is evolving rapidly and the competition between creating and detecting fakes are constantly playing catch up.   

Source: CNBC

Post a Comment