Meta CEO Mark Zuckerberg on June 29 at a major computer design conference stated that his company would be using generative AI to create unreal content by reacting instantaneously to a user, a development that was first reported by The AI Journal in April.
“In the future, [a user’s feed] is going to be created with these tools,” he said. “Some of it, I think, eventually is going to be content that’s either created on the fly for you, or kind of pulled together and synthesized through different things that are out there.”
Targeting a user’s vulnerabilities with AI—confirmed
In an article on April 18, “Is Facebook using AI to create false content?” The AI Journal found evidence that Meta already seemed to be experimenting with this approach—tracking a user’s interests through data about mouse hovers and internet searches, then synthesizing unreal images that would appeal most strongly to the user’s primitive brain chemistry.
During his comments, which came during the panel, “AI and The Next Computing Platforms,” with NVIDIA CEO Jensen Huang at the SIGGRAPH 2024 conference, Zuckerberg described the development as part of a “journey.”
It involved starting with providing content based simply on friends’ activities.
“We’ve kind of been on this journey where [the feed and recommendation systems have] gone from just being about connecting with your friends,” he said. “If someone did something really important like your cousin had a baby or something, it’s like, you want that at the top.”
“Breaking society”
But the voyage has been much more arduous—at least for Facebook (now Meta)—than Zuckerberg’s comments seemed to imply.
According to a book by Jeff Horwitz, a technology reporter for the Wall Street Journal, “Broken Code,” Facebook was involved—and in many cases instrumental—in not just facilitating but inadvertently promoting some of the most horrific events of the past decade.
Facebook, according to company employees interviewed by Horwitz, actively shared content leading to genocide in Myanmar. It was behind the fulmination that led to the storming of the U.S. Capitol on January 6, 2021. It was behind the selling of sex slaves and indentured servants in the Middle East. It was behind widespread abuses in India, where it courted powerholders as they rained violence and slaughter down on Muslims and others. Before taking preventative action, it turned anti-vaxxers’ comments into a firestorm that changed users’ perceptions about the COVID-19 vaccine.
Instagram, also owned by Meta, was discovered by Meta’s researchers to wholly distort girl’s perceptions of themselves, feeding anorexia-related material to the most vulnerable.
While researchers in Meta knew of all of these abuses, little was done, according to the interviews collected by Horwitz, and then usually only when there was media or political pressure.
The journey seemed largely one of ripping apart the social fabric—here and around the world, according to the book.
A new “net” for users
But, speaking at the June conference, Zuckerberg characterized the concerns of Meta, at least partially, as being concerned about not provoking the ire of its users.
“You’d be pretty angry at us if we, you know, it was buried somewhere down your feed,” he said, referring to the company’s earlier way of deciding what material a user saw, in this case, personal news from a friend or relative.
But, he said, the company’s system for “ranking” which posts and what material would appear at the top of the user’s screen was changing with the advent of AI.
Now, it appears, the system works differently—and potentially even more potently—than in the past.
The next phase of the journey involves enabling “creators” or AI tools to choose from any amount of material anywhere on the internet and use that to feed a user with real or artificially constructed content, said Zuckerberg.
The AI Journal has chronicled what appears to be Facebook already using AI to cast out a net that encompasses the entire internet, looking for something that will most entice and entrap the user’s interests—then matching it with what it already knows and is still learning about the user’s deepest, darkest, and most secret desires and dreams.
Zuckerberg hailed the new approach.
“But now, over the last few years, it’s gotten to a point where more of that stuff [the content fed to the user] is just different public content that’s out there. The recommendation systems are super important because now instead of just a few hundred or thousand potential candidate posts from friends, there are millions of pieces of content and that turns into a really interesting recommendation problem,” he said.
Soon most content on Instagram will be fed by the AI fire
Zuckerberg singled out Instagram, a platform that many users have come away from feeling personally and morally degraded, according to research cited in “Broken Code.”
And now AI will take even more control of what a user sees.
“The majority of the content that you see today on Instagram” will be drawn from the internet and not necessarily from people a user follows, said Zuckerberg.
Meta has been hit with legal action about the impact of Instagram on young people. Critics say use of it leads to self-destructive behavior—that the platform distorts reality for young people, generating shame and self-loathing.
Zuckerberg indicated that the next milestone in the company’s shifting use of AI was to create wholly unreal content, either by simply making it up or synthesizing existing elements.
The AI Journal article of April 18 described what appeared to be such a simulation.
