The Rise of Sadistic Online Exploitation

****Trigger warning - self-harm****

In May 2025, I sat down and wrote about a boy in a bedroom in Texas.

His name was Bradley Cadenhead. He liked Minecraft. He watched gore. And somewhere in the space between the two, he had been jailed after building a Discord server called 764 — a place where cruelty was not hidden but celebrated, where children were lured into video chats and told to cut themselves, undress, perform, or else. I called that piece The Children of 764. I wrote it because almost no adult I met had heard any of this, and the few who had did not yet have the words for what they were looking at. Bradley Chance Cadenhead was a 15-year-old high-school dropout from Stephenville, Texas, when he founded 764 on Discord in 2021, taking the name from the first three digits of his local ZIP code. After a house search on 25 August 2021, he was arrested, and Discord banned his accounts the same year. On 16 May 2023, in Erath County, Texas, he pleaded guilty to nine counts of possession of child sexual abuse material and was sentenced to 80 years in state prison, where he remains in the custody of the Texas Department of Criminal Justice. He was denied parole in 2025, and in February 2026 a federal judge rejected his bid to overturn the sentence. But the network did not end with him. The US Department of Justice has since described 764 as a terror network, and the FBI now treats it as a tier-one investigative priority, a decentralised ideology that has outlived its teenage founder and spread across platforms and borders under dozens of different names.

I wrote "The Children of 764" as a warning. I did not expect it to read, a year later, like a field note.

Last month, Thorn, which is one of the most serious child-protection research organisations in the world, published Advancing Response to Sadistic Online Exploitation in Networked Youth Environments. It draws on interviews with more than thirty practitioners across eight countries. It is careful, sober, evidence-led. And reading it, I felt something I have learned to distrust in this work and that is the cold relief of being right. Everything I described from a two-regional school halls in regional NSW and a handful of frightened phone calls, Thorn has now mapped at the level of the system. These 764 and the COM-style networks, the field now calls sadistic online exploitation, or SOE.

When I wrote about 764, I described networks that did not behave like the predators we had spent two decades teaching students and parents to watch for. The danger was embedded in the exact platforms our children already live inside, the games and servers where belonging is currency. The people running these networks looked ordinary because they were increasingly children, too.

I wrote about the deliberate hunt for the vulnerable and the lonely, the isolated, the kids carrying trauma or sitting in mental-health vulnerability and identity communities where their pain was visible from the outside. I wrote about coercion dressed up as love. About a girl who was told to stab herself on a livestream. About a high-achieving student in Connecticut who was groomed into becoming an accomplice and was, at the same time, a victim. The duality that now defines so many of these cases. I wrote about two girls, twelve and thirteen, in two separate regional New South Wales towns, who met their abusers through online gaming and were slowly led into Discord and into harm. Both of them, their parents told me, had said these people were their family now. One case only surfaced because a grandparent pulled me aside after a parent session in a room of fewer than twenty people, and said, quietly, that they thought it might be happening in their own home. They were right.

At the time, some people heard all this as teenage drama inflated into panic. Thorn's report is, among other things, the end of that excuse.

The most useful thing the report does is explain why our systems keep failing even as awareness grows. It is not, mostly, a failure of will. It is a failure of shape.

SOE networks, Thorn finds, work as accelerated conditioning environments. They do not groom slowly over months, the way we were trained to expect. They reach for shock early, reset a child's sense of what is normal, and build status on the severity of the harm, to themselves, to others, to animals, that a young person can be made to produce and broadcast. The cruelty is the content. The content is the belonging.

These networks, the report says, move horizontally across systems that respond vertically. A platform sees a fragment of self-harm. A reporting line sees a fragment of imagery. A child-protection unit sees a case here; a counter-terrorism unit sees a case there; a cybercrime team sees something else again. Each is looking at one limb of the same beast, and no one is mandated to see the whole. There is no shared governance to stitch the picture together, and the quiet pull of every institution is to stay inside its own lane, because that is what the institution is funded and measured to do. So the harm slips through the seams, not because nobody is watching, but because everybody is watching a different brick of the same wall.

It would be easy to read a report like this and feel that awareness work made up of the blogs, the parent nights, the warnings, was the small first act before the serious people arrived. I want to argue the opposite. The frontline noticing came first for a reason. It is where the pattern shows itself before the data can catch it. But noticing is no longer enough on its own, and I will not pretend it is. The behavioural signals I have been begging caregivers to take seriously, like sudden secrecy, an intense attachment to someone never met, withdrawal, exposure to gore, unexplained injuries, a child sliding into obscure servers, need to be understood in context by people trained to read them. Educators, clinicians and youth workers should treat digital exploitation as part of their first assessment, not as a footnote discovered three crises later.

Thorn's recommendations point exactly where my own work has been dragging me for years, away from reacting to single cases, and toward dismantling the system that keeps producing them. The report sets out three shifts that I will explain plainly:

See it sooner. We have to stop waiting for the online crime to announce itself and start reading the behaviour that comes before it — the secrecy, the sudden devotion to someone never met, the quiet slide into closed servers. In practice, that means training the adults already standing closest to children — teachers, GPs, school counsellors, youth workers — to recognise these patterns and to ask the second question, so that a withdrawn, frightened child is screened for digital coercion the way we already screen for abuse at home. It means schools running these indicators into their existing safeguarding and wellbeing processes rather than treating "online safety" as a separate, lesser category. And it means putting challenges back into the gateways the open gaming and social platforms where recruitment begins through real age assurance, private-by-default settings for minors, serious restrictions on adult-to-child contact, and reporting routes a frightened twelve-year-old can actually find and use. Right now, finding a vulnerable child on the internet is the easiest thing to do. That is a design choice, and design choices can be reversed.

Respond as one system, not five. A child caught in these networks is, in the same moment, a child-protection case, an extremism case and a cybercrime case. Today, each of those sits in a different building, under a different mandate, staring at a different part of the same harm. The how here is unglamorous but decisive. Shared assessment frameworks that hold all three lenses in one hand; standing cross-agency case conferences so police, child protection, schools and platforms are reading the same child's file at the same table and a single coordinating function with the authority to convene them, because coordination that depends on goodwill collapses the moment a case crosses a border. On the platform side, it means detection tuned to the coordinated behaviour of a network, the shared scripts, the handoffs between accounts, the escalation patterns, not just the lone piece of content that happens to trip a filter.

Build a way out. You cannot pull a child out of a fire and leave them standing in the cold. These networks give a young person belonging, status, the feeling of being seen, and until we can offer something real in its place, "just walk away" will never be an instruction a lonely child can follow. The how is borrowed from work we already know functions in extremism and gang exit through a credible mentor, a route back into a safe community, trauma-informed mental-health care, and patience measured in months, not meetings. Exit has to be designed and resourced, not assumed.

And we have to care for the people left standing in the wreckage. The investigators, educators, clinicians and families who survive this carry it long after the case is closed. That means real supervision, rotation off the worst material, and proper psychological support built into the job rather than offered as an afterthought. This work leaves deep trauma in its wake, and a system that spends its people without protecting them will not hold.

I am returning to The Children of 764 deliberately, not out of vanity, but because the line between then and now is the whole point. What looked, in 2025, like an extreme and unrepeatable horror was the early edge of something systemic. The warnings handed to parents and frontline professionals have since been echoed by the FBI, the RCMP, the AFP-led ACCCE, and now by Thorn's research. This was never just another online-safety topic to file alongside screen time and scams. It is a multi-year, multi-sector harm system that has set up camp exactly where our children now live.

I wrote, at the end of that first piece, that the future will judge us not by how we innovated but by what we tolerated — and that right now we are tolerating too much. A year and a serious body of research later, I would only add this, we can no longer say we did not know. The foreseeable risk exists now. What remains is whether we are willing to act on it before the next twelve-year-old decides the people hurting her are her family.

References

Pendergast, K. (2025). The Children of 764. Safe on Social. https://www.safeonsocial.com/post/the-children-of-764

Thorn. (2026). Advancing Response to Sadistic Online Exploitation in Networked Youth Environments. https://www.thorn.org/research/library/sadistic-online-exploitation/

Australian Centre to Counter Child Exploitation (ACCCE). Sadistic Online Exploitation. https://www.accce.gov.au/help-support/sadistic-online-exploitation