If you are worried about a child right now, contact the police on 000 in an emergency. Kids Helpline is on 1800 55 1800 and Lifeline on 13 11 14. Reports can be made to the Australian Centre to Counter Child Exploitation at accce.gov.au/report.
The Australian Centre to Counter Child Exploitation has published guidance on a category of harm it calls sadistic online exploitation. Its definition is the deliberate and malicious use of the internet to harm, manipulate or exert control over others for personal pleasure, gratification, sense of power, ideological beliefs, or to corrupt victims and inflict ongoing harm.
The distinguishing feature is the motive. In the sextortion most people have heard of, the offender wants money, and once that is understood the interaction is at least legible. Here the offender wants the harm itself, and standing within a group that rewards it. The Australian Federal Police put it directly when it first warned about this in September 2024: unlike sextortion, these offenders are not motivated by financial gain, and are instead driven by exploiting vulnerable victims into producing abhorrent content for their amusement.
That difference matters operationally. There is no payment demand to refuse, no transaction to trace, and no point at which compliance ends the ordeal. Compliance is the product.
Most sextortion is financial, and that is why this is hard to see
It is worth being straight about the proportions. The Australian Institute of Criminology surveyed 1,953 Australians aged 16 to 18 in mid-2024 and asked the 220 who had been sexually extorted what the offender demanded. Half received a financial demand: 41.4% money, 13.6% a gift card, 7.3% something else of value. Another 43.6% were asked for more intimate material. Respondents could pick more than one, so those overlap. The offence most people picture is the more common one, and nothing here suggests otherwise.
Now look at what is missing. The survey has no category for a demand to self-harm, to hurt an animal, or to perform humiliation on camera. Those fall into an unbroken-out "other" at 14.6%, or they are not captured at all. The authors are careful to say that determining offender motive was beyond the scope of the survey.
So Australia’s national dataset on sexual extortion records what was demanded rather than why, and its instrument contains no box for the category three law enforcement agencies are now warning about. That is the point worth holding onto. Sadistic online exploitation does not look rare because the evidence says it is rare. It looks rare because the counting has not caught up with it.
What authorities are describing
The pattern published by the ACCCE, the AFP and the FBI is consistent across jurisdictions. An offender builds a trusting or romantic relationship, frequently posing as a peer. Contact begins on gaming platforms, social media apps and messaging services. Once trust exists, the victim is coerced into producing content, and that content becomes leverage. From there the demands escalate, toward self-harm, animal cruelty, and in the most severe cases suicide.
The FBI has also documented doxing and swatting used as instruments of control, and the orchestration of self-harm while a victim is livestreaming. The AFP has noted that some offenders are the same age as their victims.
None of the agencies publish the mechanics in operational detail, and this briefing does not either. What they do publish, deliberately, is what an adult should be able to notice.
The scale is not marginal
The AFP announced a dedicated taskforce in October 2025 to pursue the decentralised network behind much of this offending in Australia. At announcement it had identified 59 alleged offenders, made three arrests domestically and nine internationally, and interrupted multiple cases of victimisation. The Commissioner described members as loosely affiliated sub-groups of young males from English-speaking countries, and said victims were predominantly young girls and vulnerable youth.
In February 2026 the FBI Boston field office warned publicly that the Bureau was investigating more than 350 subjects, with all 56 field offices involved. It described victims as typically underage females between the ages of 10 and 17, though anyone can be targeted, and offenders as frequently males under the age of 25. In July 2026 a Europol-led operation across nine countries referred 4,340 URLs for removal.
The AFP has asked media not to name the network by the label it uses for itself, on the basis that notoriety is part of what its members seek. That request is honoured here, and it is worth understanding why it exists. For these groups, publicity is not a side effect. It is a recruitment mechanism.
The warning signs are physical
This is the part of the ACCCE guidance that most deserves circulation, because it is unusually concrete. Unlike most online harm, this one leaves marks.
The signs the ACCCE lists are unexplained self-harm including carvings, bite marks and burns; changes to eating and sleeping; withdrawal from family, friends and activities; secretive online behaviour or a marked increase in device use; and the sudden adoption of extreme language or symbols. The FBI adds unexpected contact from unknown numbers, and any talk of suicide.
A carving is not a subtle indicator. If a parent, teacher, coach or clinician sees one and knows what it can mean, that is a genuine intervention point. Most people do not know, which is precisely why the ACCCE published the list.
Where detection fits
Everything above happens inside private conversation. It begins in a direct message, moves to a channel the family cannot see, and by the time a physical sign is visible to an adult the coercion has been running for some time.
The signals are structural rather than lexical. A relationship that forms quickly and moves private. Escalating demands. A pattern of compliance and further demand. Content used as leverage against the person who produced it. These are behavioural shapes, and they are detectable as shapes, without an adult reading a child’s messages and without waiting for the child to disclose.
Guardii detects that structure inside the direct message and group chat layer, and surfaces it to a parent or a safeguarding lead as a pattern rather than a transcript. A child in this situation has usually been convinced that disclosure will destroy them. Detection that does not depend on them finding the words is the only kind that reaches them in time.
If any of the signs above are present, contact police on 000 in an emergency, or report to the ACCCE at accce.gov.au/report. Kids Helpline on 1800 55 1800 is free, confidential and available to young people at any hour.