Well, the US Presidential Election 2020 is still some time away, but Facebook is leaving no stones unturned in maintaining the sanctity of its platform. It has imposed a ban on all sorts of manipulative videos as well as photos, also denoted as “deepfakes.” Techniques like artificial intelligence can be used for video editing. Apparently, this could completely malign a particular video, rendering it fake. Facebook believes that videos of this kind if circulated across its platform, could spread rumors and unwanted information. Thus, the social media giant wishes to put a full stop to such kind of disruptive videos by seeking a ban on the same.
Facebook announced this policy change on Monday wherein it stated that it plans to stop all kinds of fake videos from making its rounds on its platform. This decision has been taken after the controversial video featuring Nancy Pelosi, the House Speaker, was deceptively edited and leaked out last year. Apparently, the video was edited in a manner that slowed down and distorted Pelosi’s speech, which apparently gave the viewers a feeling that she was drunk. Some video-editing techniques were used upon this video in order to give this kind of effect. This video has been viewed on Facebook by many viewers across the globe.
The entire Nancy Pelosi video made an appearance on Facebook first. During that time, the social media platform had denied deleting the video because a spokesperson had stated there is as such, no policy which states that all information appearing on Facebook is true to the core.
However, this current policy would not be having any effect on satirical videos or any of those involving any parody. Nor would this new rule in any manner affect the Nancy Pelosi edited video that has already been doing its rounds. These new changes have been decided before the House Energy and Commerce Committee is scheduled to hear out a plea with regards to fake videos on Wednesday. Monika Bickert, Facebook’s Vice President with regards to global policy management, would be representing the social media giant in the scheduled hearing this week.
The proposed policy also would not be having any effect on videos where the footage has not been labeled properly or where the dialogues have been slightly tampered with. Taking account of all these factors, Hany Farid, a forensic expert at the University of California, expressed displeasure at the manner in which Facebook is tackling these “deepfake” videos. Facebook was associated with Farid’s lab when it came down to handling these “deepfakes.” Farid feels that the approach taken by Facebook was “narrowly construed.”
Farid felt that although the videos that have been ignored by Facebook did not use any AI techniques to make them manipulative, the fact remains that these videos are also misleading. Though these were not deepfakes, they made it look like the leader spoke something else, which, in reality, he or she did not. The point to be noted herein all these videos is that the videos were made to look in a manner in which they shouldn’t be. He has stated in an email that, “Why focus only on deepfakes and not the broader issue of intentionally misleading videos?”
However, the policy does ban all videos concerning women who have been pulled into pornographic videos without their consent. Such kind of videos does add a lot to online harassment. According to the research company Deeptrace Labs, these videos accounted for about 96% of the total deepfake videos last year.
During the previous year, Facebook, along with a few other companies, had organized a “deepfake detection challenge.” A reward in the form of money was offered to those who came out with good techniques to spot any kind of manipulative videos. Many such manipulative and fake videos were sent to the researchers last month. The challenge is expected to go until March 2020.
Image Source: WashingtonPost