Facebook recently said it’ll be removing all “deepfake” videos — videos that are altered by AI and intended to mislead or trick viewers. But how can Facebook actually enforce this?
Across the web, AI technology is being used to alter videos in order to spread false information. Facebook looks to be on the move to try and stop this. In recent years, social media sites, like Facebook, have been criticized for a lack of responsibility and action concerning content spreading conspiracy theories and fake information. Now, Facebook is going to specifically target deepfakes — another name for videos altered by artificial intelligence. These videos can make it appear that someone, say the President, is saying or doing something, when in fact that action never actually occurred.
Facebook says it will remove deepfakes that “would likely mislead someone into thinking that a subject of the video said words that they did not actually say.” The use of this type of video will also be banned in ads.
Will this make a difference?
When it comes down to it, it isn’t clear how much this initiative will actually prevent the spread of false information on the social media platform. Usually, videos that spread false information are edited in more traditional ways. These would consist of cutting out context or changing the order of words to push a narrative. Facebook will be putting all videos through its system for fact-checking potentially deceptive content. Content that is factually incorrect appears less prominently on the site’s news feed and is labeled as false.
Essentially, Facebook is trying to ban deepfakes before they become much more widespread. In many ways, it needs too because so many groups (lawmakers, academics, and political campaigns) are furious with how Facebook is handling the spread of false information.
Facebook won’t be attacking as many videos as you think
The new policy Facebook plans to enforce will not attack videos classified as “Cheapfakes” and “Shallowfakes” — content that is edited with a simple video-editing software. Facebook right now will only go after videos that are generated by artificial intelligence. However Facebook says they will take on deepfakes and that is still a step in the right direction. Computer scientists have said that AI-generated images and sounds can be indistinguishable from real life. The technology could substantially increase the volume of false and misleading information online if left unchecked. Also, deepfakes are becoming ever more common online.
How can Facebook enforce this?
Tech companies are researching new techniques to detect deepfakes and how to stop their spread. However, there’s currently no clear way to combat them. How Facebook plans to detect and remove deceiving deepfakes from their platform isn’t clear either. Additionally, there’s a question of what constitutes deceiving. Facebook will have to determine what is true and what isn’t, and judging by its track record, its judgment is shaky.