MISHAWAKA – Facebook’s recent ban of deepfakes, a type of video editing which transforms the subject into something unrecognizable, is a step towards combatting misinformation for the upcoming 2020 election. However, policymakers are concerned the ban doesn’t cover all of the necessary bases.
Deepfakes, a portmanteau of “deep learning” and “fakes,” are the heavily edited videos that can seemingly transform what a person is saying and doing. In May of 2019, a video of the Speaker of the House of Representatives, Nancy Pelosi, was being circulated online; in the video, Pelosi appeared to be inebriated onstage while answering a question. Upon later comparison, it was revealed the video had been altered from the original; the speed and pitch of her voice had been changed.
Such an example is now referred to as a “cheapfake,” contrasting it with the much more subtle deepfake. What makes the deepfakes so difficult to detect is the complexity of the process in creating them, namely, the use of artificial intelligence software. This is where the “deep learning” of the name appears, the learning that such a software must do to work so well. All that the aspiring editors must do in order to use some of these programs is select a video, choose any photos they wish, and insert them into the software, which will then superimpose the photos on the video in an almost indiscernible manner.
Naturally, this misinformation being widely circulated could cause disruption. The potential for sinister individuals to exploit such strategies and misrepresent others in the future prompted many to examine social media outlets’ methods of discerning fact from forgery. That scrutiny has led to Facebook’s recent banning of the deepfakes. In a blog post from Monika Bickert, Facebook’s Vice President of Global Policy Management, they clarified the requirements for taking down such content.
“Going forward, we will remove misleading manipulated media if it meets the following criteria:
- It has been edited or synthesized – beyond adjustments for clarity or quality – in ways that aren’t apparent to an average person and would likely mislead someone into thinking that a subject of the video said words that they did not actually say. And:
- It is the product of artificial intelligence or machine learning that merges, replaces, or superimposes content onto a video, making it appear to be authentic.”
As far as the efficacy of the ban, Bethel’s Assistant Professor of Religion and Philosophy and Professor of Primetime, Keith Koteskey, weighs in.
“I, probably at this point would say, tentatively, I would feel comfortable with where they’re at, even with the acknowledgment that there will be some material that is posted that is inaccurate or false,” Koteskey said. “If it seems outlandish, then that ought to at least give us pause, if it’s a case of something that just seems unbelievable.”
When on Facebook, it’s encouraged to keep in mind that this ban does not cover material that is satirical in nature or obvious parody. Furthermore, the people working to take down the misinformation aren’t perfect and not everything that is false will disappear. This is a process; a discerning eye for all users is a necessity today.
Quote courtesy of https://about.fb.com/news/2020/01/enforcing-against-manipulated-media/
Other information courtesy of The Washington Post: