'Deepfakes' in the Courtroom

32 Pages Posted: 10 Jan 2023

Date Written: October 1, 2020


Seeing is believing — but for how long? At present, people attach a lot of probative weight to images and videos. They’re taken at face value as evidence that an event occurred as alleged. The advent of so-called “deepfake” videos might change that. Thanks to advances in artificial intelligence, it is now possible to create a genuine-looking video that makes real people appear to do and say things they never did or said. Software for creating deepfake images, video, and audio is already freely available online and fairly easy to use. As the technology rapidly advances, it will become harder for humans and computers alike to tell a fake video from a real one.

Inevitably, deepfakes will start coming up in the courtroom context. This Article surveys the ramifications of deepfakes for pre-trial and trial practice, including authentication of evidence, professional responsibility, and a potential “reverse CSI effect” on juries primed to question even authentic evidence in an era of disinformation and “fake news.” Fortunately, courts are no stranger to the phenomenon of evidence tampering and forgery. The rules of evidence have long imposed authentication requirements to help screen out fakes. I argue that those requirements are sufficient as-is to deal with deepfakes, and that raising the bar for authenticating video evidence would do more harm than good. Although it may prove costly, courts will be able to handle the challenges posed by deepfakes as they have ably handled previous generations of inauthentic evidence.

Keywords: deepfakes, deep fakes, evidence, rule 901

Suggested Citation

Pfefferkorn, Riana, 'Deepfakes' in the Courtroom (October 1, 2020). Boston University Public Interest Law Journal, Vol. 29, No. 2, 2020, Available at SSRN: https://ssrn.com/abstract=4321140

Riana Pfefferkorn (Contact Author)

Stanford Internet Observatory ( email )

Stanford, CA 94305
United States

Do you have a job opening that you would like to promote on SSRN?

Paper statistics

Abstract Views
PlumX Metrics