Deepfake videos utilize artificial intelligence and machine learning algorithms to superimpose the faces of famous people, or even ordinary folks, onto the bodies of other people in different videos. Deepfake videos appear very realistic because machine learning allows the algorithm to continuously improve the image. This technology is often used in a sinister way to incriminate people in pornographic videos or to generate visual ‘evidence’ of events that never actually happened. The open-source artificial intelligence tool, TenserFlow, has been misused as a tool for creating Deepfake videos as the program relies on machine learning and image processing.

“Deepfake technology does its own google image search and scours through social media and can, by itself, replace faces in videos.  The program improves itself independently through machine learning.  Using this tech, anyone can create fake videos including pornographic videos of just about anyone.  Also, Deepfakes can be spread rapidly considering how quickly media is consumed and reproduced online. There is another technology that lies in this same vein, such as tech that can automatically alter images, and tech that can recreate voices.  The implications of this are that we may be looking at a future where people can create photos, videos, and audio of someone doing things they never did and saying things they never said.”

Development of crafty fake videos techniques influences jurisdictive decisions in courts as well. Videos and images have been strong evidence to prove eyewitnesses, but now it came to the situation that the validity of videos and images may be skeptical. Rather, videos might need to be proved by eyewitnesses. Another problem is that, yet, there are not many experts who know well about artificial intelligence and are capable of distinguishing fake videos from the real videos.

While it is only a matter of time before fake videos keep improving their performance, research in this field is sparse and so too are the necessary tools that could help stop the perpetuation of deep fake videos. Researchers have begun to make algorithms that can make these deep fake videos more easily identified but this same technology can also further advance the creation of these videos.

Sean Lee: Third paragraph, editing, and posting

Zaria: First paragraph

Gray: Second paragraph

TJ: Fourth paragraph