Read on Twitter
THREAD: (1/9) Millions of people encountered a Deepfake for the first time during the Last Dance on ESPN when State Farm ran a commercial using similar tech. While harmless the clip points to a much darker future. (1/x) https://www.nytimes.com/2020/04/22/business/media/espn-kenny-mayne-state-farm-commercial.html?smid=tw-share
(2/9) If it’s impossible to quickly disprove a video clip then people will choose whether or not to believe it based on their own pre-held beliefs. Because of this, malign deepfakes have the ability to chip away at any shared set of facts.
(3/9) Any video (or audio for that matter) that challenges my pre-held notions becomes a deepfake and any one that confirms them has to be real. If you you disagree with me, my "proof" becomes a deepfake.
(4/9) Don't believe me? Or think I'm being overdramatic? Think of how hard it is now to have a conversation with a family member who doesn't trust the "so called Mainstream Media". Now add into that conversation "irrefutable evidence". Thanksgiving just got a whole lote more fun
(5/9) And that moves us to one of the central problems deepfakes pose. Publicizing them doesn't alleviate the problem, it worsens it. The more people who know about deepfakes, the more likely people are to claim something is a deepfake with little or no evidence.
(6/9) This applies to politicians too. The "Liar's Dividend" (a phrase coined by @daniellecitron and @BobbyChesney) describes a situation where a politician claims incriminating audio or video is a deepfake even when it's not.
(7/9) Their supporters latch onto the statement as proof the audio or video isn't real, and the politician dodges a disaster as it takes weeks or months to determine whether the video was a deepfake or not. By then the narrative has changed and the world has moved on.
(8/9) So, while the State Farm/ESPN commercial was fun, it moves us one step closer to a future where we don't know whether that video we saw on twitter was real or not (if were not there already).
(9/9) Want more information about deepfakes, what they are, how we got here, potential solutions (spoiler: there aren't and won't be any time soon), and what we do next? Check out this (slightly dated) paper below. (Ya, some shameless self promotion here) https://jsis.washington.edu/news/deep-fakes-fake-news-and-what-comes-next/
