Mark Zuckerberg deepfake: Facebook refuses to bite at the hook

Copyright of the image
Bill Posters

Facebook said it would not delete a video of its leader, Mark Zuckerberg, of Instagram, in which he seems to credit his success to a secret organization.

The clip is a "deepfake", made by an artificial intelligence software that uses pictures of a person to create a video of it in action.

Facebook had already been blamed for failing to remove a tampered clip from US House of Representatives President Nancy Pelosi.

The latest decision coincided with the announcement of 500 new jobs in London.

The social network has announced that a large number of publications would participate in creating its own machine learning software to automatically detect and delete malicious content posted on its platforms.

In addition, they will create tools to help human workers examine potentially dangerous documents.

The company announced that it would create more than 3,000 jobs in the capital by the end of 2019.

Multimedia playback is not supported on your device

Mark Zuckerberg's deepfake video was created for an art installation presented in Sheffield under the name Specter. It is designed to draw attention to how people can be monitored and manipulated via social media in light of the Cambridge Analytica case, among other scandals.

He presents a computer-generated image of the chief executive's face merged with images of his body extracted from a video presentation given in 2017 in an office at Facebook's headquarters in Silicon Valley. An actor has provided the audio recording on which it is synchronized.

The 16-second clip – which is played in a loop – was uploaded to Instagram on Saturday.

However, it only became important when Motherboard reported on its existence on Tuesday.

"The result is quite realistic – if you leave the video muted," commented the news site.

"The voice superimposed on the video is clearly not Zuckerberg, but a person who is trying to print."

The account involved tagged the video with a #deepfake hashtag.

The Instagram post has been viewed more than 25,000 times. Copies were also shared via Facebook itself.

"We will treat this content in the same way that we handle all the wrong information on Instagram," said a spokesperson for the parent company's application, Facebook.

"If third-party fact checkers flag it as false, we will filter it from Instagram's recommendation surfaces such as Explorer and hashtag pages."

The artists involved said they "welcomed" Facebook's decision, while questioning the company's ethic.

"We believe that by using the art to explore this type of technology with a critical mind, we are trying to question the power of these new forms of computer propaganda and that, therefore, we are not going anywhere. We would not like to see our works censored by Facebook "BBC.

"However, we would like significant regulation and oversight of the digital influence sector."

Political row

If Facebook had chosen to block the post, he could have faced accusations of hypocrisy after refusing to remove a manipulated clip from Ms. Pelosi three weeks ago.

This video was not a pretense, but had been slowed down in places for the Democratic leader's speech to appear blurred.

The technology company said at the time that the information published on its website did not have to be "true". But he said that it would limit the frequency of posting the video in members' news feeds and provide a link to fact-checking sites.

Ms. Pelosi later criticized the firm stating, "At the moment, they are publishing something that they know to be wrong."

"I can take it … But [Facebook is] lie to the public ".

The Washington Post has since reported that Zuckerberg tried to contact Ms. Pelosi personally to discuss the issue, but she did not respond.