The Ethics of Deepfakes: Why We Should Worry and What We Can Do
Journalists, and various news outlets alike, have recently been getting a bad rep almost completely due to the birth of “fake news”. A new study actually revealed 45% of people claim to have lost trust in news media citing “misleading/inaccurate/fake news” as the top reason for their lack of trust. In the digital age, it’s getting increasingly easier for anyone to fabricate a news article to make it look completely legitimate, but what about video? Now, I know we’re all aware of the editing software that made it look like Tom Hanks was meeting JFK in Forrest Gump, but this type of face swapping technology isn’t only found on the set of the latest Hollywood films.
Popular online tech magazine, Motherboard, first reported on the emerging technology that could seamlessly edit the faces of popular celebrities into pornographic video. This technology came to be known as deepfakes and has been garnering media attention due to the potential it has to create politically based synthetic media. One publication perfectly expresses the potentially dangerous implications of deep fake technology in stating, “You thought fake news was bad? Deepfakes are where truth goes to die”. It’s important we understand how potentially threatening this type of technology is and learn ways of fact checking content in order to hopefully restore a sense of trust in news media.
Why We Should Worry
1. Deepfake technology is available to anyone at the click of a button.
If you take a second and Google “deepfake creator” a whole host of websites show up with the sole purpose of creating faked video content. There are even videos and tutorials on how users can create more realistic believable video.
https://www.youtube.com/watch?v=H9ZfftwagTg#action=share
2. Deepfakes go viral.
A new study by Statista, reports that 70% of Americans are on social media (Statista). If a deep fake video were to go viral it could potentially mislead people on a mass scale. Often times the damage is already done even if the video is later found out to be fake due to anchoring bias, the human tendency of being influenced by the first piece of information that we hear.
3. Deepfakes are protected by the First Amendment and considered “creative” works of expression…yes, even the porn videos.
The dilemma with deepfakes is that they don’t directly violate any code of ethics in particular. There are actually no laws established against this practice. Senator Bernie Sanders is currently pushing for a bill to outlaw the use of this technology and describes it as “something that keeps the intelligence community up at night”.
However, there is no reason to lose sleep over deepfakes lurking in the dark corners of Instagram. Instead of being reactive we should be proactive and learn different resources to fact check and debunk deepfakes.
What We Can Do:
My best word of advice to the social media news junkie would be to follow your instinct. If you see a video on social media and something about it looks off. Whether it be the facial expression or mannerisms, understand it could be fake and proceed with caution. It’s always better to be safe than sorry. So, if trusting your gut isn’t enough assurance that the video on your feed isn’t fake, here are a few sources that will do it for you.
1. GFYcat
This site is leading the way in debunking deepfakes by completely banning them from their site. They’ve also created an algorithm that is blocking deep fake videos from being posted to “through a combination of AI software and human moderation”.
2. Rochester Institute of Technology
The school recently won a $100,000 grant that will help pay for the development of technology that will automatically detect deepfakes through a combination of language, audio and visual information. Users will then be shown an “integrity score” to indicate whether the contents of the video have been altered.
3. YouTube Dataviewer and InVid Browser extention.
Both of these sites allow users to paste a link to a video and “get more information about its origins, as well as pull out key frames for further inspection”.
https://citizenevidence.amnestyusa.org/(YouTube Dataviewer)
http://www.invid-project.eu/tools-and-services/invid-verification-plugin/(InVid Browser Extention)
