Why Facebook has been filled with Jesus Christs made with eggs, spaghetti or cauliflowers

“No one likes my artistic work.” “I made it with my own two hands!” “May 2024 is your best year.” They’re just a few of the Facebook captions that accompany thousands of bizarre images of Jesus Christ generated by artificial intelligence. The platform’s algorithm is promoting them, and some have gone viral, attracting huge audiences to pages that later seek to make money off of, or even scam, users.

“The images are captivating: some are beautiful, others are sensationalist and there are some that are just strange,” says Renée DiResta, researcher at Stanford University and co-author of an article that was published in March about the phenomenon titled How Spammers, Scammers and Creators Leverage AI-Generated Images on Facebook for Audience Growth. Thousands of users interact with these images, many with a simple “amen” that could indicate the presence of faithful bots.

Through the years, Facebook has refined its base mission to grow its audience. The Jesus Christs are just another resource used towards this end, easy to create and ideal for grabbing the gaze of users. “The competition for attention among page creators means they’ll use any tactic that seems to work, which now includes publishing AI-generated content that the algorithm appears to be promoting,” says DiResta.

The captions added by authors serve to promote the images’ virality. On one, a little Black boy appears to one side of a Jesus Christ made of spaghetti. Its text reads: “Thanks to all who appreciate my art.” None of the images warn that they’ve been made by machines.

DiResta’s article studies 120 Facebook pages that have each posted at least AI-generated images. One sign of their success is that these pages have an average following of nearly 130,000 users. All it takes is for one to view the images for the Facebook algorithm to begin offering more examples and new variations. EL PAÍS saw pages of Latin American users that combined these religious images with memes, animals and still lives also made with AI, interspersed with captions that said “good morning” and funny videos. One of these users, a Haitian who lives in Mexico, briefly responded to the publication’s questions about whether he had made the images, whether people really liked them and if he had realized he could make money through them: “Yes”, he answered, but didn’t want to explain further. The user had more than 7,000 followers.

Beyond the tricks of Facebook, at the heart of the debate is a more far-reaching concern: if people confuse these ridiculous images with real art, what will happen when AI-generated images become truly realistic? Nothing is for certain, but at least not everyone is getting tricked. Below one of these images, a user writes: “What marvelous sculpture make hands blessed by God, my love you are a genius, God bless you today, tomorrow, and congratulations forever from Colombia.” Another responded: “Oh lady, how can you believe this stuff, you’re so easy to fool. Don’t you realize it’s a fake image? You can’t believe everything that is posted on social media.”

The credulous woman received another response that illustrates the kind of scam based on the concept that if someone believes that these fake images are real, they’re probably more naïve and easier to swindle. “I am a French woman and it’s been more than 13 years since my husband got older [sic] in an automobile accident. After his death, I inherited my husband’s fortune. Because of my advanced age and my health, I decided to economically help people who want to receive financial assistance for their bills, since I don’t have any children to benefit from my inheritance and I don’t want the government to freeze my fortune after my death. If you are interested in this economic activity, send me a message,” the user writes, adding a link to a real WhatsApp number.

“Scammer accounts sometimes interact with credulous commentators on posts, in pages as well as groups, sometimes asking for personal information about them and offering to sell them products that don’t exist,” explains the article. Once an audience has been created for a group or page, ways of extracting benefits may vary: “We marked the pages as ‘scams’ if: one, they trick followers by robbing, buying or changing page controls; two, they give fake names, addresses or other identifying characteristics; and/or, three, they sell fake products,” details the academic article.

Some of the pages that use these tactics regularly change their names and keep their followers as they switch their focus or sell their audiences. The spam pages take advantage of the attention to bring their users to pages off of Facebook that are full of ads.

Jesus Christ is just one of the lead characters in the process. In the third quarter of 2023, one of the 20 most-viewed images on global Facebook was a kitchen with a glass floor. It had 40 million views and 1.9 interaction, according to platform’s quarterly report on its most popular content. Some of the pages that utilize this scam also use animals, little kids, cabins and landscapes, anything to get the attention of Facebook users. “Just like Picasso had his blue period, these pages will go through phases, a couple dozen snow sculptures; a couple dozen carved watermelons; a couple dozen wood carvings; a couple dozen artistically arranged plates of sushi, each with a very similar style,” says the article.

“We suspect that these high levels of interaction are, in part, driven by Facebook’s recommendation algorithm,” DiResta and her co-author Josh A. Goldstein, from Georgetown University, state in the article. Two years ago, according to a leaked Meta document, the social network changed its algorithm to be more similar to that of TikTok. The goal, according to the document, was to, “help people to find and enjoy interesting content regardless of whether it was produced by someone you are connected to or not.” The authors show how the “Suggested for you” content went —in percentage of views — from 8% in 2021 to 24% in 2023.

Even though the center of operations for these viral images is Facebook, they also exist on other platforms. “The AI-generated images are also popular on Instagram,” says DiResta. “They are on Twitter and everywhere else, but the design of Facebook and Instagram make people believe certain specific kinds of visual content and with time, build audiences.”

In February, Meta announced that it would label AI-generated content on Facebook as well as on Instagram and Threads. As of now, the promise has not been fulfilled. To evaluate potential dangers, it would be an important step. Though sometimes it can be easy to detect AI-generated images, that’s not always the case: “Even in situations where there are obvious signs, identifying them still requires great cognitive effort, and it may not be realistic to expect that of the average Facebook user who is simply looking at their timeline,” says DiResta.

Sign up for our weekly newsletter to get more English-language news coverage from EL PAÍS USA Edition