"Click This If You Love Jesus": How AI Sloperations on Facebook Are Flooding 22 Million Users with Musk Memes, Disinformation, and Culture War Propaganda' #NewReport

From Musk memes and hacked celebs to #Boomerbait and disinformation, Facebook is being overrun by algorithmic sludge and right-wing propaganda—while Meta looks the other way.

Facebook is being overrun by a sprawling network of AI-generated slop (read this great piece by Jason Koebler for more on slop), hijacked celebrity pages, and engagement-baiting fake news, much of it run from Vietnam, much of it part of co-ordinated networks. The potential reach is massive, with at least 22 million users potentially exposed. In this investigation, I expose how “AI Sloperations” and boomerbait click farms are running rampant on Facebook—deceiving and stealing the attention of users via a sprawling network of co-ordinated fake news hubs, Musk-worship meme factories, hacked celebrities, and polarizing propaganda. Welcome to the world of #ZuckSuck, where AI Sloperations and click-farmed disinfo are breaking what’s left of Facebook’s credibility.

Collage of content from the network

AI Sloperations, as opposed to just AI Slop, refer to coordinated, large-scale campaigns that deploy (often) low-quality AI-generated slop—to manipulate the attention economy. Unlike isolated spam, these operations are systematic industrialised efforts to flood platforms with emotionally provocative, politically divisive, or misleading material. They rely on both well worn exploits of social media platforms, as well as Meta’s increasing indifference to policing harmful content.

Right-wing Boomerbait refers to emotionally manipulative online content designed to exploit low digital literacy, particularly among older users. It typically features sentimental stories, patriotic slogans, religious appeals, culture war content, or moral outrage—often paired with clickbait headlines and AI-generated visuals—to provoke engagement. While not always false, boomerbait is engineered to bypass critical scrutiny, exploiting nostalgia, fear, ‘traditional’ conservative values, or guilt to drive shares and clicks. (The term isn’t to disparage boomers, merely designed to reflect the likely target audience and fact those over 65 are nearly four times likely to share fake news online).

While the operation(s) detailed here are coordinated, it is not clear if there is a client paying for propaganda, or if the people running the pages are parking them to farm engagement to make money - plumping pages for the next campaign. Fundamentally it’s an example of dysinfluence. Regardless of intent - whether financial or otherwise, the outcome is likely to influence discourse, amplify polarisation, and exploit algorithmic systems for maximum reach.

What this investigation reveals isn’t just a content crisis—it’s the result of platform failures that enable identity theft, political manipulation, and transnational deception. At the heart of it is Meta’s refusal to enforce even the most basic protections, whether that’s rapidly labelling AI slop, disinformation, or acting when people’s accounts are hacked to leverage their networks to spread disinformation.

Coordinated Facebook Disinformation and AI-Generated Slop Network

This analysis documents over 130 Facebook pages disseminating coordinated disinformation, AI-generated content, low-quality clickbait, and manipulative propaganda to at least 22,000,000 people. The pages span several thematic categories including celebrity fan pages, news & media outlets, entertainment hubs, model profiles, and military fandom. All exhibit suspicious behavior patterns indicative of inauthentic or repurposed accounts.

Key Findings

• Geographic anchoring: The vast majority of these pages appear to be operated from Vietnam (noted in over 120 entries), with frequent secondary links to the Philippines, Indonesia, Morocco, Pakistan, and the U.S. This points to a probable Vietnam-based content farm network with global targeting.

• Hijacked celebrity or public figure pages: Many verified or legacy accounts (e.g., Rolonda Watts, Andrew Gruel, Dondré Whitfield, Maggie Oliver) have been compromised and now serve up AI slop, Musk worship memes, porn-adjacent Telegram links (e.g., t.me/Xvideo18Age), or generic emotional bait. Several celebrities have publicly acknowledged the hacks and how Facebook have done nothing to stop them.

• Musk-centric and Right-Wing Propaganda: Dozens of pages—particularly those themed around U.S. patriotism, entertainment, or military fanfare—promote Musk-centric hero narratives, anti-Left culture war tropes, and clickbait headlines about veterans or celebrities. These often follow a pre-set narrative template, combining AI-generated boomerbait with partisan sentiment manipulation.

• AI and Clickbait Economy: Many pages display traits of automated content generation, such as repetitive structures, strange phrasing, and viral templates. Some even contain AI prompts. Ads linked to these pages are mostly clickbait thumbnails, hinting at low-cost international ad farming. A significant number of ads are catalogued in the Facebook Ad Library.

• Fake News Brands: Multiple "news" pages share identical content, URLs (like news.autodailyz.com), or layouts, including "Daily News Feed", "The Daily Scoop", "Newsflash Central", and "Freshy Pressed News". These mimic legitimate news aesthetics but are entirely fabricated, spreading hoaxes, fabricated quotes, or celebrity death rumors.

• Page Recycling & Thematic Shifts: Several accounts show a sharp content shift between 2023–2025, particularly around March–August 2024. Originally legitimate pages are now rebranded as fake news or Musk-worship hubs, often without any transparency updates.

• Review-Based Detection: Many accounts are flagged in user reviews as fake, AI-run, or malicious. Comments reveal user awareness of hacks, often requesting Facebook intervention or warning others.

• Tied Networks & Shared Infrastructure: Shared domains (autodailyz.com, cafex.biz), common Telegram links, reused ad templates, and similar follower counts indicate these pages are likely part of a larger coordinated influence operation or engagement farming network, possibly monetized via affiliate traffic, ad revenue, or data collection.

The Grift

For those of you still able to stomach Facebook (which in theory is many, but perhaps not Zuckerberg, who acknowledged that social media is dead), you may have come across highly shared memes mentioning something great Musk has done, or maybe something salacious the Pope and/or Whoopi Goldberg said - if only you’d click the link in the comments. I first came across this when my Facebook feed would consistently serve me up articles about Musk with thousands of likes, shares and comments. Frequently, the images were AI-generated, with often-times completely fabricated stories about how he had done something amazing. (I have no idea why I was being served this content - especially given I think Musk is as likeable as chlamydia).

Some examples of the AI-slop

The stories range from the comically bizarre to the somewhat sinister. There are posts reporting that Musk is training squirrels to make Tesla batteries, Musk inserting robot legs on dogs, Musk adopting pets, Musk adopting an endless stream of children, and Musk feuding with Imane Khelif. The below photo grid is a screenshot from a typical page. You can see that the pages devote a lot of time to posting about Musk.

The posts would frequently go viral. For example, below is a story about Musk-as-whistle-blower, threatening to name those involved in the Diddy scandal. Of course it’s not true, Reuters, Snopes and others have had to fact check it. Numerous accounts shared identical content and graphics, at times the number of shares was in the tens of thousands.

Example viral posts showing the reach of the false posts

My personal favourites though, stories of Elon Musk adopting dozens of children, seemingly from around the world. (Entirely false of course but with seismic engagement) Check out the posts below.

Creepy AI images of Musk adopting/stealing children?

Numerous viral and false posts have claimed Musk has adopted children from around the world

Also this photo of Xi Jinping Putin and on BYD hover scooters is pretty epic.

Fact-checking organisations often debunk these stories, but don’t tend to emphasise the fact that these campaigns are often co-ordinated and connected, playing instead a game of endless whack-a-mole - pointing to the sheer impossibility of keeping up with all the falsehoods.

The pages roughly fell into two categories, pages that had seemingly been constructed specifically for this kind of operation, and those that had been hacked and repurposed for the operation. A subset of the pages seemed to mimic ‘breaking news’ accounts, while others were sports pages, film pages, and models. The total number of followers is estimated to be over 22,000,000 (These are people who have liked the FB pages). Of course it is not clear how many have been exposed to the content. It could be much higher, or lower. The below chart is a list of all the pages, along with the number of followers, arranged by category.

Table of pages by followers

Shared links

Many of the pages shared identical links, such as URLs to a telegram channel featuring scantily clad women, or to clickbait websites that may be crypto scams (need to look into these more). As the diagram below shows, different pages were often linked by the fact they were sharing the same links (in addition to similar AI Slop).

The Asia Connection

Some of the pages were new, but many had been hacked or appropriated or repurposed. Page transparency details showed that almost all 130 of them were run from Vietnam, Philippines and/or Indonesia, with a portion from Macedonia. Many were run from multiple countries, but Vietnam was by the far most common country to be mentioned. As many readers will know, Vietnam is a hotspot for mercenary content farming operations and scams. Seem for example this investigation by myself and

Sohan Dsouza

into how how one network itself turned itself to virulent anti-immigrant propaganda. There’s also a new ARTE documentary that mentions this. You can watch it here.

Pseudo-News Sites

A subset of the pages had profile photos that imitated the professional graphics of mainstream news sites, with logos and slogans such as ‘breaking news’. No doubt these sites were trying to give the illusion to the casual viewer that they were legitimate news sites. Below is a collage of profile photos from 23 of these pages.

These pages are very active, posting anything from 4 - 15 times per day (sometimes more). While most of the posts get a moderate amount of engagement, occasionally they will go viral. Whether this virality is a product of statistics (i.e. if you post enough eventually one will go viral), or other manipulation is unclear, although many of the pages are also running ads (more on that later) to increase engagement. It’s worth remembering that a similar type of pseudo-news site (Channel 3 Now News) was key in propagating the fake news about the Southport killer back in 2024. The man who ran that was briefly arrested in Pakistan.

Additionally, there is a persistent use of pseudo-news aesthetics—headlines like “BREAKING,” “LIVE,” and “SHOCKING”—that simulate the appearance of journalism while avoiding the burden of factual reporting. This visual language amplifies the sense of urgency, further blurring the line between actual news and manipulative content.

Switching identity

After looking at a number of these pages, it was quite clear they were all sharing identical content, operating in batches. For example, a few of these pages, including Insightful Updates, Fresh News Today, the Daily Buzz, all switched from tweeting Musk AI slop to content about country music, and then movies. They also changed their overall identity but kept the same name. When they switched to posting about country music, they changed their profile picture from news sites to music-related images. Below are some examples of how this looked.

The pages would change their identity at the same time, and to the same topic

Coordination

Looking at the limited ad data where available allows us to sometimes see 'the ‘beneficiary’, or the page/account that runs the ad campaigns. For example, you can see below the account ‘Flash News Feed’ (now changed to ‘Popcorn Pilgrims’) has an ad run a beneficiary called ‘1’.

This tree visualization shows how a large number of these pages—many posing as independent news outlets or public figures—are being centrally managed by a single administrator - an account labeled “1.” This account alone controls at least 17 different pages. There are a number of other examples where multiple pages are being controlled by the same entity. This data was extracted via the ad library. Because of the limitations on transparency, it is likely the rest of the network follow the same structure.

Tree graph showing ‘beneficiaries’ of FB pages

Further evidence of the links between the pages is demonstrated in the below diagram. This shows that the beneficiary page ‘1’, was also running pages that were promoting different links. Again, this suggest a larger operation was coordinating multiple entities promoting different URLs. It also highlights the fact FB can see this co-ordination, but choose to do littler about it.

Musk Love

Focusing on the Musk content, which accounts for many of the viral posts, the general thrust is to portray him in a positive light. I did a visual analysis of around 150 posts from these Musk-focused pages and grouped the themes in the table below. The content is sort of a Musk-glamour-slop blend of algorithmic exploitation, emotional manipulation, disinformation and aesthetic fantasy. Musk is generally portrayed as a benevolent father, rebellious genius, and philanthropic savior. (There’s also a lot of pictures of AI-augmented women wearing very little clothing).

‘Type 2’ News Sites

Another subset of pages all appeared to also be marketing themselves as news, or current affairs websites. As you can see from the profile pictures and names, the person running this dimension of the network appears to have the same graphic designer. My personal favourite has to be ‘adulting affairs’.

Collage of profile pictures of a subset of the network

Coordination

In addition, many of these pages were running ads, some of which were flagged for inappropriate content, allowing an insight into the beneficiary and providing evidence of central coordination. What became clear is that the same beneficiary was running most of these type 2 news pages, establishing a definitive link. Below you can see that ‘Finlife’, ‘VN’, and ‘Thien Loc’ were running multiple pages. It is probable that the same entity was also running those three beneficiary pages. Unfortunately META’s transparency tools are limited so it is hard to know more, but we can at least confirm via metadata the coordination of multiple fake news pages.

The Ads

Many of the Type 2 network were running ads. The ad techniques are common to every page in the whole dataset that was running ads, not just the Type 2 network (another indicator of central coordination). The pages are running hundreds, probably thousands of ads that all follow a similar formula. These ads are classic examples of "boomerbait" — emotionally manipulative, low-effort visual content designed to exploit users’ beliefs and values to drive engagement, particularly likes, shares, and comments. Each ad uses a strong emotional or moral appeal (religion, patriotism, admiration for Elon Musk) alongside a visual prompt to "Press this thumb," clearly intended to provoke a reflexive response (also indicate non-native English).

Phrases like "If you love Jesus" or "If you believe in God" are paired with generic AI-generated or stock imagery of Jesus. The 'thumb’ icon and directive ("Press this thumb") simplify the action to a single click, turning faith into an engagement tool. Themes of patriotism; "If you respect veterans, press this thumb" taps into patriotic duty and respect for military service, particularly appealing to older or conservative users. The Musk variant ("If you're a fan of this guy") places Musk among religious or heroic figures, or perhaps just relies on his position as polarizing figure.

These kinds of manipulative engagement traps aren’t just harmless sentimentality. At best, they exploit vulnerable users for cheap clicks. At worst, they’re part of a larger strategy to grow disinformation networks by plumping up FB pages, monetize emotional manipulation, and redirect users toward scams, propaganda, or AI-generated slop. That Meta continues to allow — and even algorithmically reward — this kind of content speaks to continued failures of platform governance.

From Trump love to emotional manipulation

The general thrust of most of the posts within the ‘Type 2’ subnetwork are emotionally manipulative, politically slanted, and engagement-driven. The images and headlines reveal a consistent set of themes and narratives designed to provoke strong emotional responses—anger, pride, sorrow, or nostalgia—primarily among older Facebook users.

A dominant theme is the lionization of Donald Trump, who is repeatedly framed as a heroic, embattled, or wronged figure. Headlines portray him as a patriot facing persecution or as a triumphant disrupter of corrupt elites, often through exaggerated or fabricated “breaking news” announcements. There is a lot of content celebrating military service and law enforcement, particularly white veterans and uniformed officers. These figures are depicted as moral anchors, often alongside flags, gravestones, or touching farewells, meant to evoke patriotic reverence and a sense of national nostalgia. There is also content disparaging Pride, and trans people. Although a handful of posts may appear ideologically neutral or even left-leaning (such as reposts of content by Occupy Democrats), the overwhelming thrust of the content is right-leaning, both visually and rhetorically.

Another major theme is sentimental tragedy, particularly around death, illness, and sacrifice. Thumbnails highlighting cancer diagnoses, fallen paramedics, or elderly patients "finally going home" are common. These stories blend grief with redemptive framing to maximize emotional impact and viral potential. Often, these are paired with boomerbait tropes: sad puppies, birthday signs held by veterans, children reunited with parents—stories designed to guilt the viewer into sharing.

Some of the AI-generated bait is bizarre. Take this example, hundreds of permutations of an AI-generated image of a crying girl who painted a picture of her deceased father, but no one “supported her”. You can see below multiple images, all quite similar.

My favourite is the image where the dad has cats on his shoulders.

For the love of god won’t someone support her!

And you think no ones engages with this, wrong. They often get thousands of reacts, shares and comments. Interspersed among the emotional appeals are posts invoking culture war anxieties. These include veiled or explicit swipes at Democrats, trans individuals, Michelle Obama, and liberal elites. Prior to the US election many were putting out provocative statements pitting immigrants against veterans, or not-so-veiled misleading statements about immigrants, English proficiency, and US citizenship (English is actually required for US citizenship unless you have an exemption). Below are four posts from different pages (About Life, Amazing Story, Happiness Mount) sharing almost identical content.

Celebrity culture is also exploited, with posts about aging icons, musicians’ deaths, or glamorous figures like Melania Trump. These function as low-effort attention bait, often detached from reality, but often effective at pulling in clicks and shares through name recognition and sentimentality. Finally, several thumbnails deploy religious or inspirational tones. Letters to God, redemptive endings, or quasi-spiritual moral tales are peppered throughout.

Taken together, the visual and textual patterns in this dataset reveal a coherent media style designed to inflame and polarize. This is algorithmic propaganda wrapped in the aesthetics of sincerity, producing high engagement. The network exploits affective vulnerabilities, particularly among demographics susceptible to nostalgia, nationalism, and moral panic.

While the majority of content appears to lean right, it’s not always the case - which is not uncommon if the aim is engagement and polarization. At times the network is more ambiguous. One page went from posting Musk-glamour-slop to posting AI generated billboards with messages not consistent with one particular party. Indeed, they were posting content both celebrating Trump’s America while also saying it was like 1930s Germany.

Celebrity Hacks - Some Examples

Another dimension of this network is the hacking of public figures' social media accounts, which are repurposed to push viral disinformation or engagement bait. The shocking thing about this is how poor Facebook are at addressing it. It also indicates how these hacks are rarely one off opportunist hits, but the result of seemingly organised and co-ordinated repurposing of accounts.

List of hacked celebrities/public figures or athletes with verified accounts

Andrew Gruel - An American chef and television personality, Andrew Gruel is the founder of Slapfish Restaurant Group and has appeared on Food Network’s Food Truck Face Off. Dondré T. Whitfield - An American actor known for roles in The Cosby Show and All My Children, Dondré T. Whitfield has received three Daytime Emmy nominations. Katie O - An 18-year-old country singer from Jacksonville, Florida, Katie O gained recognition as a contestant on Season 26 of The Voice, representing Team Reba. Rolonda Watts - An Emmy Award-winning announcer and actress, Rolonda Watts is known for hosting her self-titled talk show and for her work in television, radio, and film. David Benoit -A contemporary jazz pianist and composer, David Benoit has released over 40 solo recordings and is a five-time Grammy nominee. Mamma Bruschetta - A Brazilian actress and television presenter, Mamma Bruschetta is known for her characters Condessa Giovanna and ET Zero, and has appeared in various theater productions. Kokane - Born Jerry B. Long Jr., Kokane is an American rapper from Pomona, California, recognized for his collaborations with N.W.A and Above the Law. V. Bozeman - An R&B singer and actress, V. Bozeman gained fame for her performance of "What Is Love" on FOX's Empire and is a Grammy-nominated artist. Angela Dimitriou - A Greek pop-folk singer, Angela Dimitriou is renowned in Greece and the Middle East, with hits like "Margarites" topping charts in countries like Lebanon. Te'a Cooper - An American professional basketball player, Te'a Cooper played as a guard for the Los Angeles Sparks in the WNBA and had a notable college career at Baylor. Jake Shields - An American mixed martial artist, Jake Shields has competed in organizations like the UFC and Strikeforce, known for his grappling skills and championship titles.

Jake Shields - verified

Jake Shields is an American mixed martial artist who has competed in the UFC, Strikeforce, and EliteXC. He is known for his background in wrestling and Brazilian Jiu-Jitsu, and increasingly, his far-right politics. His Facebook page of 481000 followers was hacked, with the new admins ostensibly running it from Vietnam. They posted a lot of clickbait, AI slop about Musk.

Jake Shields noted in February that his page was hacked and he’s been unable to contact anyone at Facebook to fix it.

Rolonda Watts - verified

Rolonda Watts is an Emmy Award-winning announcer and the promotional voice of SHERRI. She has acted in television series including Mind Your Business and Survival of the Thickest. Her work spans multiple media platforms, including acting, podcasting, comedy, and writing.

Rolonda Watts’ Facebook page appears to have been compromised and was being used to spread low-quality clickbait. Recent posts falsely attribute philanthropic acts to Elon Musk, accompanied by AI-generated images and misleading headlines.

Rolonda Watts has publicly confirmed that her official Facebook page was hacked and has been used to post fake news, racist propaganda, and misleading clickbait. Despite being a Meta-verified public figure, Watts stated on X (formerly Twitter) on March 12, 2025, that she has been reporting the issue for over eight months without any resolution. The hacked page has continued to share AI-generated posts falsely linking celebrities like Elon Musk and Ronda Rousey to charitable causes, as well as other forms of "boomerbait" designed to drive engagement. Watts has criticized Meta’s inaction, describing it as negligence that enables identity theft, trademark abuse, and reputational harm.

Maggie Oliver - verified

Maggie Oliver is a former British detective who became a whistleblower after exposing failures in the investigation of the Rochdale child sex abuse ring. She later founded the Maggie Oliver Foundation to support survivors of abuse. As of 07/05/2025 her fb is still hacked, with transparency showing it’s been run from Vietnam. Lots of Musk and now ‘wellness’ content. Reviews noting this, “don't even think it's her page anymore....looks hacked to me.” One user even called out how triggering it was to see this kind of content on her page.

Andrew Gruel - verified

Andrew Gruel is an American chef, restaurateur, and media personality known for founding Slapfish seafood restaurants.

Gruel publicly acknowledged the issue, stating that he has reported the page to Facebook multiple times without success. Despite being verified and widely followed, the fake account continues to operate, monetizing engagement without consent. Gruel sarcastically remarked that he just wants “his cut,” and urged others to report the impersonator.

Meta seem uninterested

In addition its seeming indifference to plight of numerous hacked pages, Meta had said it would label AI-generated content, but this seems to depend largely on people labelling it or metadata. If you’re a bad actor using genAI there doesn’t seem to be much stopping you from posting it on the platform without any warning.

Another interesting element is the inconsistency with which META applies content warnings. In some cases, posts from the network have been flagged as false, but on other occasions they haven’t, despite being fact-checked. Below you can see a viral post shared by much of the network about Musk paying for a Neuralink implant for a sick girl. This was fact checked by Snopes and others, and FB issued a false information warning.

Another viral post about Musk threatening to withdraw support for a boxing organisation over Algerian boxer Imane Khelif was also fact checked and found to be false, and the posts are still all over FB with no fact check.

Who are they kidding?

Of course it would be easy to dismiss this absurd content as pointless, meaningless, or ineffective. But that would be naïve, and indeed, wrong. Sure, there are a lot of people obviously identify the content as fake, or AI-generated. I really liked this comment left in the Review section on one of the pages:

““Conservative disinformation, hate and other right wing dreck mixed in with lots of pseudo-patriotic dressing. Probably AI-generated, courtesy of Mark Zuckerberg's cesspool of disinformation. The unwitting American masses are lapping it up and spreading it amongst themselves.”

However, the issue is not that some, or even many people might be able to identify the dysinfluence. The issue is that so many people cannot. The below screenshot of a viral post (104,000 reacts and over 11,000 shares!) is telling. Not only are there two AI-images of Musk with dogs, the top comments appear to be extremely credulous and complimentary of Musk - with no indication that they think there is anything untoward about the post.

Conclusion

This investigation has exposed a sprawling, coordinated disinformation network operating across at least 130 Facebook pages with a combined reach of over 22 million users. There are likely hundreds more, perhaps thousands reached. If one considers the potential reach it is staggering - 130 pages publishing an average of 6 posts per day for five months would produce around 117,00 posts.

Run primarily from Vietnam and neighboring countries, the operation leverages hijacked celebrity pages, AI-generated clickbait (“slop”), right-wing propaganda, and sentimental "boomerbait" to flood the platform with emotionally manipulative and politically polarizing content.

At the center of the operation is a set of administrator accounts—especially one labeled “1”and ‘VN’—that centrally control dozens of seemingly independent news or public figure pages. Many of these pages impersonate legitimate news aesthetics or have been repurposed from formerly authentic accounts. Others push Musk-centric mythologies, Trump hero worship, and pseudo-patriotic stories designed to manipulate older users through grief, pride, nostalgia, or moral outrage. In general, the content is polarizing or propagandistic - or both.

Facebook's ad library and transparency tools reveal recurring ad sponsors, shared infrastructure, and identical content templates, pointing to a highly coordinated sloperation - although as with most big tech platforms, these transparency tools are wholly inadequate. Despite repeated public complaints from verified celebrities like Jake Shields, Rolonda Watts, and Andrew Gruel—whose pages were hijacked—Meta has taken little meaningful action. This is not just a content problem. It is a systemic platform failure. Facebook is enabling identity theft, mass deception, and the monetization of misinformation at scale. In a very real sense, we are seeing the gruesome convergence of relaxed content moderation, AI slop, post-truth politics, and clickbait farms.

For more on origins of AI slop read this great piece on 404 by Jason Koebler.