The Lazarus Network: The Dead Follower Syndrome

THE LAZARUS NETWORK

The Dead Follower Syndrome

PART 1: THE EMPTY AUDIENCE

You have followers you have never met.

This is not an observation. It is not a complaint about social media. It is a statement of fact so universal that you have stopped questioning it. You have followers you have never met, whose profile pictures you have never examined, whose usernames you have never read aloud, whose existence you have accepted the way you accept furniture in a room you walk through every day.

They are there. They have always been there. You do not know when they arrived.

I want you to do something right now. Not later. Not after this video. Right now. Open your phone. Go to your follower list. Not your following list — your follower list. The people who chose to see your content. Scroll past the names you recognize. Past your friends. Past your family. Past the accounts you vaguely remember following you back after you followed them.

Keep scrolling.

You will find them in the middle. Not at the top — those are recent. Not at the bottom — those are old friends. In the middle. A cluster of accounts that share a specific set of characteristics so consistent that once you see the pattern, you will not be able to unsee it.

The profile picture is a real photograph. Not AI-generated — real. A real person in a real location with real lighting and real imperfections. The kind of photograph that was taken between two thousand twelve and two thousand eighteen, when smartphone cameras were good enough to be clear but not good enough to be cinematic.

The bio contains exactly three to five emoji. A hobby. A relationship status or family reference. A single inspirational word or phrase. The bio reads like it was written by a human being. Because it was. Once.

The account follows between eight hundred and fifteen hundred other accounts. It has between two hundred and a thousand followers of its own. It has posted between eight and thirty times. The posts are photographs — meals, sunsets, pets, children at birthday parties, a vacation beach.

And the last post is between three and ten years old.

In two thousand twenty four, a cybersecurity research team at the University of Amsterdam published a paper that received almost no mainstream coverage. The paper was titled "Coordinated Inauthentic Persistence: Dormant Account Networks and Post-Mortem Digital Activity." The title alone should have made headlines. It did not.

The Amsterdam team had developed a behavioral clustering algorithm that could identify coordinated account networks not by what the accounts posted, but by the temporal pattern of their micro-interactions. Likes. Follows. Brief profile visits. The invisible actions that leave no visible trace on anyone's feed but are recorded in platform telemetry.

They analyzed eleven million accounts across three platforms over a fourteen-month period. Their algorithm identified what they called "dormancy swarms" — clusters of accounts that had stopped posting original content but continued to perform micro-interactions in synchronized patterns.

The swarms were enormous. The smallest contained eight hundred accounts. The largest contained over forty thousand. And they were coordinated with a precision that eliminated any possibility of coincidence.

Every account in the swarm had posted original content at some point. Every account had a real profile picture. Every account had a bio that read like a human being wrote it. And every account had stopped posting between three and ten years ago. Not deactivated. Not deleted. Just… stopped.

But they had not stopped interacting. The accounts continued to follow new users. Continued to like posts. Continued to perform the invisible micro-actions that social media algorithms interpret as signals of an engaged, authentic audience.

And here is the detail that made the Amsterdam researchers request additional security clearance before publishing their findings.

The accounts were not following random users. They were following specific users. Users who had recently been identified by advertising algorithms as "high-influence micro-targets" — ordinary people with small but highly engaged audiences whose purchasing decisions ripple outward through their social networks.

The dormant accounts were being aimed. Not scattered like seeds. Aimed like weapons.

Someone was paying for this. Someone was operating these swarms. Someone had access to thousands of dormant accounts with real histories, real photographs, real bios — and was deploying them in coordinated campaigns targeting specific individuals.

The Amsterdam team traced the command infrastructure through fourteen layers of proxy servers, three cryptocurrency mixing services, and a shell company registered in the Seychelles. At the end of the chain, they found a marketplace. Not on the dark web. On the regular internet. A website with a clean design, professional copy, and a pricing page.

The marketplace sold access to dormant social media accounts in bulk. The pricing was tiered by account age, follower count, and what the marketplace called the "trust coefficient."

And the product descriptions used a term that the researchers had never encountered before.

"Heritage accounts."

Heritage accounts.

The word "heritage" implies inheritance. It implies something passed down. Something left behind by someone who is no longer here to use it.

The Amsterdam researchers noted the terminology in their paper without further comment. They were cybersecurity specialists, not investigators. They documented the technical infrastructure, published their findings, and moved on to other projects.

But one member of the team did not move on. A doctoral student named Asha Mertens, who had been responsible for the manual verification phase of the research — the part where a human being actually looked at the accounts, one by one, to confirm that the algorithm's classifications were accurate.

Asha Mertens looked at four thousand two hundred accounts over the course of three months. And she noticed something that the algorithm was not designed to detect.

The profile pictures matched obituaries.

PART 2: THE NECRO-BOT HIJACK

Asha Mertens did not set out to cross-reference social media profiles with death records. She was verifying account authenticity — confirming that the profiles identified by the clustering algorithm were real accounts with real histories, not recently fabricated imitations.

But verification requires looking. And Asha Mertens was thorough.

The first match was Robert Calloway. She found his obituary on the second page of a Google search for his name and hometown, which were both visible in his social media profile. The obituary was from two thousand nineteen. His account had liked fourteen posts in the past month.

She told herself it was a coincidence. Someone with the same name. A common face. A mistake.

The second match was a woman named Patricia Huang. Died in two thousand seventeen. Her Instagram account had followed thirty-seven new users in the past quarter.

The third match was a teenager named Devon Williams. Killed in a car accident in two thousand sixteen. His Twitter account had liked a cryptocurrency promotion four days ago.

By the time Asha Mertens had cross-referenced three hundred of the four thousand two hundred accounts in her verification sample, she had confirmed forty-seven direct matches between active dormant accounts and published obituaries.

Forty-seven dead people whose social media accounts were actively engaging with the living internet.

Not in a metaphorical sense. Not in the way we say someone "lives on" through their social media presence. In the operational, technical, server-log-verified sense. These accounts were being accessed. Commands were being issued through them. They were following, liking, and in some cases commenting — generic comments, single emoji, the kind of interaction that algorithms reward but humans rarely examine.

The dead were participating in the internet. And no one had noticed because no one looks at their follower list the way Asha Mertens looked at hers.

She expanded her methodology. Instead of manually searching for obituaries, she built an automated cross-referencing tool that compared profile photographs against digitized obituary databases, memorial websites, and genealogy platforms. The tool used facial recognition — not the sophisticated real-time systems used by law enforcement, but a simple image-matching algorithm that compared profile pictures against photographs published in death notices.

She ran it against the full dataset of dormant accounts identified by the Amsterdam clustering algorithm. Eleven million accounts.

Three point two percent. Of eleven million dormant accounts identified as part of coordinated inauthentic swarms, three point two percent belonged to people who were verifiably dead.

That is three hundred and fifty-nine thousand accounts.

Three hundred and fifty-nine thousand dead people, active on social media. Following. Liking. Commenting. Shaping algorithms. Influencing what the living see, read, and believe.

And that was only the accounts Asha Mertens could verify — the ones whose obituaries were digitized and publicly accessible. The true number, she estimated in a supplementary analysis that she never published, could be between two and five times higher. Because not everyone gets an obituary. Not everyone's death notice is digitized. Not every country maintains accessible records.

The conservative estimate: three hundred and fifty-nine thousand.

The realistic estimate: over a million.

The question that Asha Mertens could not answer — the question that drove her to work eighteen-hour days for eleven weeks until her academic advisor intervened — was not how. The how was straightforward. Abandoned accounts with weak passwords, accounts linked to email addresses that were themselves abandoned after the owner's death, accounts on platforms that had no mechanism for reporting a user's death and removing their profile. The how was a failure of infrastructure. A gap in the system that no one had bothered to close because no one had realized it was a door.

The question was why.

Why specifically target dead people's accounts? Why not simply create new fake accounts, as bot farms had done for years? Why go to the trouble of identifying deceased users, gaining access to their profiles, and reanimating them?

The answer was on the pricing page of the marketplace. In the phrase that Asha Mertens would circle in red ink and pin to the center of her corkboard.

"Average trust coefficient: ninety-four point seven percent."

They are using the dead because the dead are trusted.

PART 3: THE TRUST SCORE MARKET

Every social media platform maintains a system that it does not publicly acknowledge. The terminology varies — "credibility index," "authenticity rating," "behavioral trust metric" — but the function is identical. Every account is assigned a score. The score determines how the platform treats that account's actions.

A new account — created today, with no posts, no followers, no history — has a trust score near zero. Its likes carry no weight. Its follows trigger spam filters. Its comments are shadow-suppressed. The platform treats it as guilty until proven innocent, because the platform has learned, through years of bot warfare, that new accounts are overwhelmingly fake.

An account created in two thousand twelve by a human being who used it for six years — who posted photographs of their children, who argued about politics, who left a birthday comment on their sister's wall every March, who misspelled words and used the wrong emoji and exhibited all the beautiful, chaotic inconsistency of a real human life — that account has a trust score that approaches the theoretical maximum.

It is algorithmically invisible. Its actions pass through every filter. Its likes register as genuine engagement. Its follows are counted as organic growth. Its comments appear without delay, without review, without the invisible hand of moderation touching them.

And when that human being dies, the score does not die with them.

The score persists. The account persists. The history persists. And the trust — that precious, painstakingly accumulated trust — sits there. Unguarded. Unmonitored. A vault with no lock, in a house with no owner, on a street where no one is watching.

This is the market. Not a metaphor. A literal marketplace with buyers, sellers, and a commodity that replenishes itself every time someone dies without deleting their social media accounts.

Asha Mertens' investigation eventually led her to three distinct tiers of the heritage account trade.

Tier One is the bulk market. Low-cost packages of dormant accounts sold to influencer marketing agencies, small businesses, and social media managers who need to inflate follower counts. These accounts follow, occasionally like, and never comment. They are the foot soldiers — the background noise of artificial engagement. A package of five hundred costs less than three hundred dollars. The buyers rarely ask where the accounts come from. The sellers never volunteer the information.

Tier Two is the amplification market. Mid-range packages of high-trust dormant accounts sold to political campaigns, cryptocurrency promoters, and disinformation networks. These accounts do not merely follow — they engage. They like specific posts at specific times to trigger algorithmic amplification. They follow specific users to manipulate recommendation algorithms. A coordinated action by two thousand heritage accounts with trust scores above ninety can push a post from obscurity to a platform's trending feed in under four hours.

Tier Three is the one that Asha Mertens almost did not include in her research because she was not certain anyone would believe her.

Tier Three is the identity market. Individual heritage accounts — not bulk, not packages, but single accounts — sold to buyers who need a specific type of digital identity. A middle-aged woman from the Midwest. A college student from London. A retired engineer from São Paulo. The buyer specifies the demographic, the location, the age range, the interests. The seller delivers a real account, with a real history, belonging to a real person who is really dead.

The price for a Tier Three account ranges from two thousand to fifteen thousand dollars, depending on the account's age, engagement history, and the completeness of the deceased owner's digital footprint.

Fifteen thousand dollars for a dead person's identity. Not their Social Security number. Not their bank account. Their social media presence. Their digital face. Their accumulated trust.

And the buyers at Tier Three are not marketers. They are not political operatives. They are not influencer agencies.

They are AI training networks.

The most sophisticated large language models — the ones that generate text, analyze sentiment, produce content that is indistinguishable from human writing — are trained partially on social media data. The models learn what human communication looks like by studying billions of examples of human communication.

But as the internet has filled with synthetic content — AI-generated text, bot interactions, machine-produced engagement — the training data has become contaminated. Models trained on contaminated data produce contaminated output. The industry calls this "model collapse" — a recursive degradation where AI trained on AI output becomes progressively less human with each generation.

The solution, for certain operators, is to ensure that training data comes from verified human sources. And the most verified human sources on the internet are the accounts with the highest trust scores. The accounts that platforms have spent years confirming are real, authentic, and human.

The accounts of the dead.

The dead are training the machines that will speak for the living.

PART 4: THE DIGITAL SÉANCE

Her name is Linda Ortega. She is fifty-three years old. She lives in a two-bedroom apartment in Albuquerque, New Mexico, with a tabby cat named Professor and a refrigerator covered in photographs held up by magnets from places she has visited with her son.

Her son's name was Marcus. He was twenty-four when he died. Acute lymphoblastic leukemia. The diagnosis came in January of two thousand twenty. The treatment lasted eleven months. Marcus died on December second, two thousand twenty, in a hospital room with white walls and a window that faced the parking lot.

Marcus had an Instagram account. He posted photographs of sunsets, his friends, his cat before Professor — a calico named Doctor who died two years before Marcus did. His last post was from September of two thousand twenty. A sunset photographed from his hospital room window. The caption read: "Not bad for a Tuesday."

Not bad for a Tuesday.

After Marcus died, Linda did not touch his account. She did not delete it. She did not memorialize it. She did not even log in. The account existed as Marcus had left it — a small, honest archive of a young man who liked sunsets and cats and had a dry sense of humor about dying.

Linda sometimes opened Instagram and looked at Marcus's profile. Not every day. Some weeks, not at all. But when she did, she scrolled through his posts the way you might turn the pages of a photo album. Slowly. With the kind of attention that only grief can produce.

On March fifteenth, two thousand twenty-four — three years and three months after Marcus died — Linda received a notification on her phone.

Marcus_sunsets liked a post.

Linda tapped the notification. Instagram opened. The activity log showed that marcus_sunsets had liked a sponsored post from an energy drink brand called VoltRush. The post was a photograph of a muscular man running on a beach with the caption "Fuel Your Fire 🔥 #VoltRush #Energy #NeverStop."

Marcus — her Marcus, who spent his last months too weak to walk to the bathroom without help, who joked about sunsets because he was not sure how many more he would see — had liked a post about fueling your fire. About never stopping.

The algorithm did not know it was being cruel. The algorithm does not know anything. It was executing a task. A heritage account designated marcus_sunsets had been assigned to a Tier Two amplification campaign for a beverage company's product launch. The campaign required twelve thousand likes from high-trust accounts within a six-hour window. Marcus's account — trust score ninety-three point four, created two thousand seventeen, last original post two thousand twenty, no red flags, no irregularities — was one of twelve thousand accounts activated for the campaign.

Linda Ortega reported the account. She clicked "Report," selected "This account may be hacked," filled out the form, and submitted it. She received an automated response within thirty seconds: "Thanks for your report. We'll review this and take action if we find a violation of our Community Guidelines."

Three weeks later, the account was still active. Still liking. Still following. Still performing.

She reported it again. Same automated response. Same result.

She tried to recover the account — to log in as Marcus, to change the password, to do anything to make it stop. But the email address linked to Marcus's account was his university email, which had been deactivated six months after his death. The recovery process required access to that email. Without it, the platform's security system — the same system designed to prevent unauthorized access — prevented Linda from reaching her own son's account.

The system that could not stop a bot network from operating Marcus's account could very effectively stop his mother from shutting it down.

She contacted support. She waited fifteen business days. She received a response requesting a death certificate. She mailed a death certificate. She waited twenty-two more business days. She received a response saying the death certificate had been received and the case was "under review."

During those thirty-seven business days, marcus_sunsets liked eighty-four posts, followed nineteen new accounts, and commented on three posts with single emoji — a fire emoji, a heart emoji, and a thumbs up emoji.

Eighty-four likes. Nineteen follows. Three comments. In the voice of her dead son. While she waited for a corporation to acknowledge that he was dead.

On day forty-one, the account was finally memorialized. The word "Remembering" was added before Marcus's name. The profile was locked. No more likes. No more follows. No more comments.

But Linda Ortega does not use the word "memorialized." In the interview she gave to a local Albuquerque news station — an interview that was aired once, at eleven PM, between a weather report and a used car advertisement — she used a different word.

She said they held his account hostage.

She said the internet made her son work after he died.

She said she had to prove he was dead to a machine that already knew he was dead and did not care.

Linda Ortega's story is not unique. It is not even rare. A two thousand twenty-five survey conducted by the Digital Legacy Alliance — a nonprofit organization that advocates for posthumous digital rights — found that fourteen percent of respondents who had lost a family member in the past five years had observed unexpected activity on the deceased person's social media accounts.

Fourteen percent.

One in seven grieving families. Watching their dead interact with a world that has moved on without them. Watching algorithms puppeteer the digital remains of the people they loved. Watching and being unable to stop it because the systems designed to protect accounts from unauthorized access cannot distinguish between a mother trying to lay her son to rest and a hacker trying to steal his identity.

The dead have more rights on social media than the living who mourn them.

PART 5: THE 4TH WALL BREAK

I have a request.

Not a suggestion. Not a rhetorical exercise. A request that I am making of you specifically, right now, in this moment, because you have spent twenty-eight minutes understanding something that cannot be un-understood.

Pick up your phone.

Open your social media. Any platform. The one you use most. The one where you have the most followers. The one you think you know.

Go to your follower list.

Scroll past the names you recognize. Past your friends. Past your family. Past the people you actually know.

Keep scrolling.

You will find an account. Maybe more than one. An account with no profile picture, or a profile picture that was taken years ago. An account that follows eight hundred people and has forty-three followers of its own. An account that has not posted since two thousand eighteen or two thousand nineteen.

An account that watched your story yesterday at three in the morning.

They did not watch it.

The person who owned that account was buried in two thousand nineteen. Their name was Elaine. She was thirty-one. She liked hiking and terrible puns and she had a dog named Biscuit who outlived her by two years. She posted her last photograph on a Tuesday — a trail somewhere in Oregon, the light coming through the trees in columns, the caption a single word: "Breathe."

She does not breathe anymore.

But her account does. Her account follows. Her account likes. Her account watches your stories at three in the morning because the server farm in Bucharest that operates her profile runs its engagement cycles during off-peak hours when the algorithmic scrutiny is lowest.

You posted a photograph of your dinner last Tuesday. Elaine liked it. You saw the notification and did not think about it. You did not recognize the name. You did not click on the profile. You accepted the like the way you accept air — automatically, unconsciously, as a feature of the environment you inhabit.

You are performing for an audience of corpses.

Every like you have ever received may include likes from the dead. Every follower count you have ever checked includes the dead. Every metric you have ever used to measure your relevance, your reach, your value as a human being in the digital attention economy includes the dead.

The platforms know this. They have always known this. They do not remove dormant accounts because dormant accounts inflate the platform's user metrics. A platform with two billion accounts can report two billion users to advertisers, to investors, to the public. It does not matter that millions of those accounts are operated by no one. It does not matter that hundreds of thousands are operated by the dead. The number goes up. The stock price follows.

You are not the customer. You are not the product. You are the living half of an audience that includes the dead, and the platform profits from both halves equally because to an algorithm, engagement is engagement. A like is a like. A follow is a follow. It does not matter whose thumb pressed the button.

It does not matter if there was a thumb at all.

The next time you pick up your phone. The next time you check your notifications. The next time you see that someone liked your post, watched your story, followed your account.

Ask yourself one question.

Are they alive?

[Then — a single notification sound. The Instagram notification chime. Brief. Bright. Familiar.]

[Black screen. Small white text, center frame:]

user_elaine_k_1987 started following you.

# [END]

| 3 | 1 | Visual + FP | Monitor wall, counts ticking up | | 4 | 1 | Visual | Network graph — "SWARM 14" | | 5 | 1 | Visual + FP | Archery target with profiles, pulse radiating | | 6 | 1 | Visual | SocialLegacy Pro marketplace | | 7 | 2 | Visual + FP | Split screen — profile vs obituary | | 8 | 2 | Visual + FP | Corkboard investigation — 47 matches | | 9 | 2 | Visual + FP | Terminal — "CONFIRMED MATCHES: 23,847" |

| 3 | 1 | Visual + FP | Monitor wall, counts ticking up | | 4 | 1 | Visual | Network graph — "SWARM 14" | | 5 | 1 | Visual + FP | Archery target with profiles, pulse radiating | | 6 | 1 | Visual | SocialLegacy Pro marketplace | | 7 | 2 | Visual + FP | Split screen — profile vs obituary | | 8 | 2 | Visual + FP | Corkboard investigation — 47 matches | | 9 | 2 | Visual + FP | Terminal — "CONFIRMED MATCHES: 23,847" |

| 3 | 1 | Visual + FP | Monitor wall with 16 dormant profiles, counts ticking up | | 4 | 1 | Visual | Network graph — "SWARM 14 — 12,847 NODES" | | 5 | 1 | Visual + FP | Archery target with profile pictures, pulse radiating outward | | 6 | 1 | Visual | SocialLegacy Pro marketplace — "Heritage Account Solutions" | | 7 | 2 | Visual + FP | Split screen — social profile vs obituary, activity updating | | 8 | 2 | Visual + FP | Corkboard investigation — 47 confirmed matches, red strings | | 9 | 2 | Visual + FP | Terminal — "CONFIRMED MATCHES: 23,847", rain on window |