In a World of Deep Fakes..

On the Left is the Tom Cruise Impersonator, On the Right is a deepfake image of Tom. Created by Chris Ume, Image from Vice.

Deep fake AI technology allows people to create fake or misleading videos, images and soundbites of popular figures doing things they have never done or said.[i]  Deep fakes are extremely prevalent in the entertainment industry, and according to a survey, around 96% of deep fakes are used in congruence with pornography, and 99% of those Deep fakes are of women who work in the entertainment industry.[ii] The deep fake editing software is readily available online for free and consequently, there has been an emergence of legal concern surrounding its continued use. 

California’s Assembly Bill 602, which passed State Senate in 2019, added a privacy provision concerning Deep fakes to its civil code.[iii] AB-602 grants celebrities and anyone who has been non-consensually depicted in a deep fake video of sexual nature the ability to sue the creator of the video in addition to those who intentionally shared it. Many states have since banned the use of Deep fakes as it pertains to doctored images and audio regarding public figures such as politicians as well as pornography. Although the law has begun to crack down on Deep fakes involving explicit sexual material, there are many websites such as Twitter and TikTok that continue to host Deep fake content of celebrities in a non-sexual manner – and the law surrounding this use is non-existent.[iv]

A potential solution to this issue could be to either ban the use of Deep fakes through international treatises or for the platforms where Deep fakes seem the most prevalent to create filtration systems that flag, mark, and remove Deep fakes as they appear on its sites, much like Twitter’s general filtration system concerning sensitive material.[v] Intel Technology in the UK, for example, developed a software (FakeCatcher) that detects Deep fake videos with over 90% accuracy by detecting blood flow signals in the pixels of a video to determine the authenticity of the face or person in frame.[vi] This technology shows that Agencies in the U.S. can make more of a conscious effort to prevent Deep fakes. The Defense Advanced Research Projects Agency (DARPA), National Science Foundation, and Intelligence Advanced Research Projects Agency (IARPA) can continue their efforts in making detection algorithms to help improve the extent to which Deep fake technology is allowed to be utilized online.[vii]

Although it is clear that there is much work to be done surrounding the ever-growing use of deep fakes and its legal scope, there are a few concerted efforts in the US and elsewhere that may greatly influence its reduction, but only time will tell how effective these efforts may be.

[i] How a deepfake Tom Cruise on TikTok turned into a very real AI company, CNN, Rachel Metz

[ii] The State of Deepfakes, DEEPTRANCE, Henry Ajder

[iii] Assembly Bill 602, CA Legislature

[iv] Deep Fake Tom Cruise Fooled the Internet, DAILYDOT, Libby Cohen

[v] DeepFakes and the Law of Armed Conflict: Are They Legal?, LIEBER INSTITUTE, Eric Jensen

[vi]FakeCatcher can detect Deepfakes In Real Time with 96% Accuracy, METRO, Anugraha Sundaravelu

[vii] Disinformation On Steroids, CFR, Robert Chesney