Social networking platform Facebook has recently announced that it plans to remove any videos from the site that have been modified in any way by artificial intelligence in order to deceive an audience.
So-called ‘deepfakes’ are videos that have been computer-generated using AI technology to look real. For example, this could be a person being superimposed onto a video to look as though they were actually at that location.
While deepfakes are relatively uncommon on the platform, and on the internet as a whole, they are starting to become more prevalent. This has resulted in Facebook stating that these present the site, and the industry, with a significant challenge as they ‘distort reality’.
The targets of deepfake videos are often celebrities and politicians, and those behind them usually use AI software to replace, superimpose, or merge content onto a video with the aim of creating an authentic illusion.
Facebook has said that it will now be removing any videos from the platform should it be identified as being edited in a way that wasn’t obvious to the average person, or if the video was to mislead viewers into thinking that the person or people appearing in the video said anything they may not have actually said.
Facebook’s staff will team up with independent fact-checkers and will judge the authenticity of videos, removing the ones it believes to be bogus. However, the company notes that this new policy will not be applicable to satire and parody videos.
In September, the company announced that it would be contributing $10m (£7.6m) to a fund aimed at improving deepfake detection technology. The chief executive of the company, Mark Zuckerberg, has himself been featured in a deepfake video in the past, in which a computer-generated version of him credited a secretive firm for the success of Facebook.
Facebook plans to work alongside government officials, businesses and academics to expose those behind deepfakes. The social network is not the only company that is trying to combat these hoax videos, with other tech giants including Microsoft and Google also having the same aim.
The company also stated that it will continue to remove videos, including graphic violence, hate speech, nudity and voter suppression.