Meta wants to mimic TikTok: towards AI-driven disinformation

“We’re basically shifting from having most of the content that you see in Facebook and Instagram come from your friend or follow graph, to now, you know, over time, having more and more of that content just come from AI recommendations.”

It was with these words that Zuckerberg confirmed to CNBC in an interview on 22 June that the Meta social media platforms Facebook and Instagram are going to follow the path set by TikTok.

He said nothing about the importance attached to reliable information in this new role being assigned to artificial intelligence, although his Chinese rival has been much criticised for the provision of false and misleading information. As the Californian tech giant’s current model, which prioritises interactions between users, is already harming access to reliable information, RSF is alarmed by a new strategic course that is taking one of the new champions of disinformation as a model.

“The big platforms are more innovative when it comes to defending their economic interests than they are about combatting disinformation,” said Vincent Berthier, the head of RSF’s Tech Desk. “By following the path forged by TikTok, Mark Zuckerberg is transitioning to a new era of prosperity for disinformation. The response to disinformation cannot be to promote it by means of artificial intelligence.”

A video app owned by the Chinese company ByteDance, TikTok is a formidable competitor for the US platforms. It claims to have 1.5 billion users and to have seen a 215% surge in earnings in 2021. Its success is based on its main feed, called “For You,” which is powered by an artificial intelligence that draws on all the content hosted on TikTok.

Any user can therefore be exposed to any content, without any transparency with regard to the reasons for the choice of content. The entire information distribution circuit is delegated to a complex and secret machine whose effectiveness offers high advertising earnings because it manages to hold the attention of its users for a long time. And this innovation is in the process of becoming the norm.

Social media apps such as Facebook and Instagram already pose serious problems for access to quality information. Their current models mix social interactions and curation algorithms to select posts and display them on users’ newsfeeds. The role they play in distributing and ranking content on the accounts users choose to follow is already problematic because it reduces the visibility of journalistic content, as RSF has pointed out. A Washington Post article in September 2021 reported that disinformation was six times more visible than informative content during the US presidential election. 

The way TikTok works is now posing new problems. Disinformation continues to represent an economic interest for the major social media platforms. In the future, each platform could have its own disinformation universe, one tailor-made by an artificial intelligence. This has already been seen since the start of the war in Ukraine where, as in Kenya, TikTok has established itself as a new champion of disinformation.