Following Melon Musk’s acquisition and rebranding of Twitter to ‘X’ in October of 2022, policies governing safety and conduct were changed or discarded, and the number of engineers tasked specifically with “trust and safety issues” was reduced from 279 at the time of Musk’s acquisition to 55 in May 2023. 

Australia’s eSafety office issued a legal notice to X in June of 2023 under the Online Safety Act that compelled the company to provide information about how it was addressing online hate across its platform in the period 24 January 2022 to 31 May 2023 inclusive. In a report prepared by the eSafety Office based on the results of their inquiry, X disclosed that it had reinstated 6,103 Australian accounts that had previously been banned from the platform. 194 of these accounts had been initially suspended/banned for hateful conduct violations. Our CEO Dr Andre Oboler raised this issue directly with the Royal Commission on Antisemitism and Social Cohesion during his second testimony: 

Since his acquisition, Mr Musk has advanced a claim that X is a platform for ‘free speech’, and it is in accordance with this logic that accounts previously suspended for hateful conduct have been reinstated. This position has attracted critique, including during the aforementioned Royal Commission on Antisemitism and Social Cohesion, where it was argued that X’s policies represented a “double standard”. Namely, the rationale invoked by X to justify its protection of potentially harmful content, a purported desire to create an environment that permits unfettered free speech, is contradicted by the platform’s concomitant failure to enforce the standards of behaviour necessary for any such environment to emerge. 

Although some figures who were reinstated under Mr Musk, such as the neo-Nazi Andrew Anglin, have subsequently been banned again, an array of individuals known for promoting hate with considerable reach and influence have had their accounts restored and remain on the platform at the time of writing. The following are a small sample of prominent social media personalities that were exiled from Twitter but were, in effect, exonerated by and welcomed back to X.

Andrew Tate: Misogyny

The notorious misogynist Andrew Tate was banned from Twitter in 2017 shortly after he asserted that women “bear some responsibility” for being victims of rape. Tate’s account was reactivated on or around 18 November 2022. Mr Tate resumed making regular misogynistic proclamations almost immediately. Amongst a wide range of harmful generalisations and stereotypes, Mr Tate’s rhetoric also involves the dehumanisation of women. Mr Tate espouses a view of gendered power dynamics in which women exist in a fundamentally and inalienably parasitical relationship with men. According to Mr Tate, women compensate for various physical, cognitive and psychosocial impairments by engaging in exploitative relationships with men that are based on exchanges of sexual gratification and reproduction for power and wealth [see below].

In addition to being offensive, Mr Tate’s production and distribution of misogynistic content has visibly contributed to the normalisation of misogyny as well as an increase in the frequency of specifically misogynistic instances of anti-social behaviour. Moreover, Mr Tate’s posts routinely violate X’s rules and policies. Of particular note is the manner in which Mr Tate’s comments about women on X violate the subcategory of ‘dehumanization’ as it is articulated within X’s Hateful Conduct policy [see below].

Mr Tate’s remarks about women consistently invoke (albeit more or less directly) his stated belief that women are essentially and categorically interchangeable with female reproductive functions. Such remarks are demeaning and degrading, but they also necessarily involve the denial of agency, and in so doing, they are inescapably dehumanising. 

At the time of writing, Mr Tate has 10.8 million followers on X.

Laura Loomer: Islamophobia

Laura Loomer, a self-proclaimed Islamophobe and white-nationalist, was banned from Twitter in 2018 following her Islamophobic abuse of a United States congresswoman. On 8 December 2022, Loomer’s account was reinstated on X. Most recently, Ms Loomer’s anti-Iranian rhetoric has attracted the attention of the Iranian regime, who have begun posting AI-generated videos depicting Loomer as a Lego Minifigure. 

As recently as March 2026, Ms Loomer made a post on X with a video taken by a passenger aboard a flight from Nashville to Florida that was diverted to Atlanta following reports by passengers of a bomb threat [see below]. In her post on X, Ms Loomer asserts that an “Arabic looking Muslim passenger onboard the plane threatened to blow the plane up with a bomb”. Ms Loomer claims that there was no media coverage of the incident, and declares that all Muslims should be deported from the United States of America. Although the flight was diverted and a SWAT team did indeed take a passenger into custody, this proved to be the result of xenophobic assumptions on the part of the passengers rather than a thwarted instance of Islamic extremism as it was portrayed by Ms Loomer.

X users contributed context via the community notes function to Ms Loomer’s post, explaining that the suspect was speaking a foreign language, and that he was in possession of some kind of countdown on his phone [see below]. It was later realised that this was simply a timer for Ramadan prayers. 

Available information indicates that the individual never threatened anyone, and the bomb threat was introduced into the narrative as a result of xenophobic stereotypes. Although no charges were filed and the suspect was cleared by the state police as well as the FBI, Ms Loomer left the post on X unedited.

The post is demonstrative of Ms Loomer’s enduring commitment to inciting hostility towards Muslims, as well as a lack of concern for the threats presented by mis and disinformation. This type of content violates X’s current Hateful Conduct Policy, which includes a dedicated subsection for “Incitement” [see below].

At the time of writing, Ms Loomer has 1.9 million followers on X.

E. Michael Jones: Antisemitism

After being suspended from all mainstream social media platforms (including Twitter) in 2018 for his production and distribution of antisemitic propaganda, E. Michael Jones’ account was restored on 23 November 2022. One of the first public remarks made by Mr Jones following his reinstatement to X was an antisemitic ‘prayer’ for God to “break the yoke of Jewish tyranny…” [see below].

Mr Jones has maintained an antisemitic publication, Culture Wars, since 1996 (previously Fidelity, which he founded in 1981), and has espoused an exceptionally broad range of antisemitic tropes in his writing and public speaking appearances. Mr Jones’ antisemitic commentary is often couched in his ideas of Catholicism, both as a religion and an institution, against which he claims Judaism exists in hostile opposition. Below is a recent example of the type of content posted to X by Mr Jones, which in this case is in the form of Holocaust denial. Having previously held the view that the Holocaust was a “justified response to Jewish behaviour”, Mr Jones’ ideas have since evolved into Holocaust denial. 

In the above example, Mr Jones’ antisemitism is reinforced by and enmeshed with his Holocaust denial via his conspiratorial claims of “Jewish privilege” and the “Holocaust narrative”. Mr Karp refers to a lampshade that was made from the skin of Buchenwald concentration camp prisoners. The lampshade was one of a number of artifacts made out of human remains at Buchenwald, and it has become symbolic of the extreme brutality faced by victims of the Holocaust. Forensic testing has conclusively proven that the lampshade is indeed made of human skin, but Holocaust-deniers like Mr Jones nevertheless advance various pseudo-academic claims about such artifacts in their efforts to challenge the veracity of Nazi genocide.

Posting material of this kind very clearly violates, at least, the following element of X’s current Abuse and Harassment policy that specifically prohibits Holocaust denial [see below].

At the time of writing, Mr Jones has 133,500 followers on X.

Dr Jordan Peterson: Amplification

Dr Jordan Peterson, in accordance with the hateful conduct policy of the time, was banned from Twitter in June 2022 after he posted transphobic abuse. On 18 November 2022, Mr Musk personally announced Dr Peterson’s X account would be reinstated. Leveraging the perceived weight of his academic credentials, Dr Peterson’s contributions to social media are shot through with climate-denial, transphobia, misogyny, and a peculiar brand of pseudu-intellectualism. Notably, Dr Peterson has continued to post transphobic remarks to X of a very similar kind to those for which he was previously banned. Instead of targeting and abusing individuals, Dr Peterson now regularly targets transgendered people more broadly [see below].

In the above example, Dr Peterson first denies the gender identity of trans women, but then proceeds to pathologise them as ‘autogynephilic’, and then condemns them as “men of the worst sort”. Dr Peterson’s remarks, in this instance, are directed at Admiral Rachel Levine, the first openly transgendered four-star commissioned officer in any of the United States uniformed services. Although these remarks meet the general criteria of ‘hateful conduct’ as well as ‘abuse and harassment’ as they are described in X’s policies, the subsection that appears to be intended to protect transgendered people from abuse offers no protections from Dr Peterson’s comments [see below]. 

Dr Peterson’s divisive, hateful and erroneous commentary presents a myriad of risks to the public sphere, and the way in which his relationship with Melon Musk has allowed him to bypass legitimate social sanctions is demonstrative of additional hazards. In the below screenshot of a post on X, Dr Peterson expresses his gratitude to Mr Musk for posting a portion of an interview in which Dr Peterson denies (among other things) the existence of patriarchy in Western societies. Notably, Mr Musk posted this video from his personal X account, and endorsed it as “true”, thus imparting artificial credibility, and dramatically increasing the public’s exposure, to Dr Peterson’s remarks. This interaction demonstrates that the reactivation of banned accounts on X can be leveraged by Mr Musk to manipulate public discourse on X and, potentially, to distort public opinion.

At the time of writing, Dr Peterson has 6.1 million X followers.

Concluding Remarks

The examples discussed here clearly demonstrate that a range of hateful behaviours that were considered unacceptable and prohibited by Twitter are tolerated by what has now become X. Moreover, X’s policies that prohibit hateful and abusive conduct are, at most, only selectively enforced, and they therefore do not represent a genuine framework for behavioural norms and expectations on X. Individuals, such as those examined here, have now had years to probe for a boundary, and they have found that, in practice, there are none. As a result, X is now an environment in which anti-social and harmful behaviour proliferates freely.

Recent events, such as the Royal Commission as well as other incidents involving regulatory bodies from around the world have demonstrated a clear willingness on the part of X to ignore and resist practically any attempt to influence the platform’s operations. As a privately owned entity, traditional mechanisms used for enforcing compliance with international standards, such as fines, do not have the same implications for X Corp as they previously did with Twitter (which was a publicly traded company). This presents significant challenges that will need to be overcome if an intervention is to be successful. Although possible, events to date appear to indicate that unless a powerful motivating factor is introduced, reform on X remains unlikely.

Despite Mr Musk’s proclamations of being a “free speech absolutist”, X’s re-activation of banned accounts went well beyond the ostensibly commendable objective of expanding participation by virtue of eliminating exclusion. Mr Musk’s reactivation of previously banned accounts was a significant instance of top-down intervention in the public sphere that necessarily changed the distribution of sociopolitical perspectives represented on the platform. These consequences have been compounded further still by Mr Musk’s support for specific socio-political ideas, attitudes and beliefs, which has served only to further distort public opinion. Moreover, Musk’s personal endorsement of hateful or otherwise problematic socio-political discourses on X serves to amplify their reach and influence. 

In any event, and setting aside valid questions about what Mr Musk is entitled to do with his property, the most fundamental problems raised by X are as a symptom rather than a disease. Given the same opportunities and affordances as Mr Musk, it stands to reason that similar if not identical outcomes would eventuate on any other social media platform if it were to implement the same conditions as X. Although it is playing an increasingly significant role in the production and consumption of misogyny, xenophobia, bigotry and prejudice, X in and of itself cannot explain the appeal of these discourses. 

That said, one would still hope that most people would be inclined to agree that the kinds of posts authored by the likes of Mr Tate, Ms Loomer, Mr Jones and Dr Peterson inflict significant harm and offer little, if any, utility or value. Deactivating at least some of the accounts that were reinstated under Mr Musk is not a curative remedy for antisocial behaviour, but it still promises better outcomes than doing nothing.

Join us on social media

OHPI LinkedIn • OHPI Instagram • OHPI Facebook 

The post “Free Speech” Selectively Enforced on X first appeared on Online Hate Prevention Institute.

Article: https://www.australianjewishnews.com/moving-forward-together/amp/

The post AJN: Progressive Judaism’s Biennial Conference to feature OHPI first appeared on Online Hate Prevention Institute.

The hybrid event was attended by 70 people. It involved a keynote from our CEO, Dr Andre Oboler, followed by a panel discussion in which he was joined by Iman Atta OBE (Director of Tell MAMA), and Prof Ali Hessami (IEEE Society for the Social Implication of Technology Chapter Chair for UK and Ireland). Q&A with the panel and audience followed.

The event examined antisemitism, anti-Muslim hate, and online misogyny and the implications for national security as hostile actors promote hate to weaken democracies. Dr Oboler’s presentation includes a look at data from three of OHPI’s reports:

• Online Antisemitism After October 7
• Online Anti-Muslim Hate and Racism Against Palestinians and Arabs: October 2023 – February 2024
• Technology Facilitated Violence Against Women

The event took place in the stunning National Liberal Club, hosted by the Defence & Security Circle chaired by Noel Hadjimichael FRSA, who also served as our very capable MC for the evening.

The post London: Defeating Online Hate first appeared on Online Hate Prevention Institute.

“The women detailed the abuse they received in a report released on Wednesday, while the Online Hate Prevention Institute (OHPI) was also commissioned to write a report on their experience.

The report found the targeted abuse was fuelled by misogyny and disinformation that tech platforms were unwilling or unable to stop.”

Patrick Scolyer-Gray, author of the OHPI report, told supporters gathered at Australia’s Parliament House on Wednesday that the case of Collective Shout was internationally significant.

“Platform moderation systems proved inadequate, disinformation spread faster than corrections, abusive material remained online long enough to create the appearance of public consensus.”

The OHPI report made 47 recommendations, the first being that harmful and sexist content specific to women should be identified immediately.

Read the ABC article here.

Read the OHPI report, “Technology Facilitated Violence Against Women”, here.

The post Threats of rape, murder and deepfake porn used against women who challenged abusive video games first appeared on Online Hate Prevention Institute.

Listen to the program here. (4m20s)

Read the OHPI report, “Technology Facilitated Violence Against Women”, here.

The post High profile women threatened by online abuse first appeared on Online Hate Prevention Institute.

“The report, from the Online Hate Prevention Institute, details a year-long campaign of hate against Collective Shout, an all-female organisation that challenges the objectification of women and the sexualisation of girls in media, advertising and popular culture. The group has suffered continual misogynistic online abuse, threats and misinformation since 2025.

The report suggests that the structures supposed to protect people online from this type of organised, anonymous, global campaign, including police, the eSafety Commission, social platforms and other gatekeepers, are inadequate.”

Read the ICDA article.

Read the OHPI report, “Technology Facilitated Violence Against Women”, here.

The post New report warns authorities not equipped to deal with coordinated online hate first appeared on Online Hate Prevention Institute.

This report examines the sustained campaign of hostility, harassment and harm directed at Collective Shout during 2025-26 following its advocacy to remove video games depicting rape, incest, sexual torture and child sexual abuse material from the digital distribution platforms Steam and Itch.io. The investigation finds that the backlash extended far beyond a policy dispute. Instead, it constituted a large-scale, multi-vector, coordinated campaign of abuse conducted across multiple online environments and characterised by misogyny, disinformation and harassment and attempts to undermine the organisation’s staff and operational capacity.

The post Technology Facilitated Violence Against Women first appeared on Online Hate Prevention Institute.

This briefing includes themes of animal cruelty, child exploitation, sexual abuse, and suicide. If you have seen disturbing content online or are having distressing thoughts, you can reach Lifeline on 13 11 14 or Beyond Blue on 1300 22 4636.

The Australian Federal Police has published a guide to Understanding Sadistic Online Exploitation for parents and carers, which can be accessed here.

Australia’s eSafety Commissioner has issued a Reporting Steps flowchart for what to do when you see distressing content online, which can be accessed here.

For several years, law enforcement agencies globally have warned about the emergence of online extremist networks that target vulnerable individuals – primarily minors – and coerce them into acts of sexual humiliation, animal cruelty, violence, and self-harm.

The dangers posed by these networks are now being scrutinised in two concurrent court proceedings in different parts of the world. Both cases are examining accountability for the tragic death by suicide of Jay Taylor, a 13 year-old transgender boy from Seattle who was targeted on Discord by a leader of the online network known as 764.

In January 2026, a German court began hearing the case against Shahriar J., a 21-year-old German-Iranian medical student who operated under the alias ‘White Tiger’. He faces 204 charges involving more than 30 victims, including murder in connection with Jay’s death and five counts of attempted murder.

The following month, Jay’s parents filed a wrongful death lawsuit in the United States against the communication platform Discord.

The facts of this case (below) are deeply disturbing and raise urgent questions about responsibility for online abuse, as well as what governments, police agencies, and technology platforms can do to prevent similar tragedies from occurring in the future. 

The tragic death of Jay Taylor

On 16 January 2022, Jay made a post on Discord looking for gay or transgender friends who shared his interest in craft projects. He was contacted by a 12-year-old girl located in Finland who had been subjected to months of coercion, blackmail, and escalating pressure to self-harm by White Tiger. When she expressed her desire to stay alive, she was told by White Tiger that in order to do so, she would have to find somebody else to die in her place.

Under White Tiger’s direction and control, the girl began communicating with Jay. She identified Jay as a vulnerable target given his history of anorexia, depression and suicidal ideation, and pressured him to enter a joint suicide pact. Impersonating the girl, White Tiger provided detailed instructions on how to commit the act while on a video call with other 764 network members. White Tiger encouraged Jay to live-stream his suicide on Instagram which took place in the early morning of 17 January, only one day after he had begun his search on Discord for friendship and community.

Jay’s parents initially had no knowledge of the online events preceding his death. They learned what had happened after a teenage girl in Australia entered a chat in which White Tiger and others were sharing the recording of his death. The girl, who had herself been targeted by 764, located Jay’s father and sent him the video.

Jay’s case demonstrates one of the most disturbing features of these online networks: that children who have been groomed and abused themselves can be coerced into recruiting and manipulating new victims. It highlights the cross-border nature of the network, with victims, offenders and witnesses located in different countries, making detection, investigation and prosecution more difficult. It also shows how abuse can move rapidly across platforms – from initial contact and coordination on Discord to livestreaming of violence on Instagram.

The 764 network

764 is a decentralised, global network that operates within a broader online ecosystem known as ‘the Com’ (short for ‘the Community’). 

The Com consists of overlapping online spaces that are associated with nihilistic violent extremism (NVE), a form of extremism in which violence, cruelty, and social disruption are treated as ends in themselves rather than as tools for achieving a political objective. Elements within the Com have been linked to accelerationism, an ideology which encourages acts of chaos and destruction as a means to bring about the downfall of society. Some parts of the network use neo-Nazi, white supremacist, and other extremist imagery.

The original 764 group was founded during the COVID pandemic by Bradley Cadenhead, then a 15-year-old living in Texas, who named it after the first three digits of his local postcode. Cadenhead adopted his tactics of grooming and extortion from another Discord community called CVLT. After Cadenhead’s arrest in 2021, 764 split into several groups that retained similar imagery (below), methods and goals. New figures emerged within the network, including White Tiger.

764 typically recruits participants and targets victims on online platforms where young individuals play games, socialise or seek companionship – such as Discord, Telegram, Roblox, Minecraft, Twitch and Steam. Members deliberately target children with mental health challenges or marginalised identities, including LGBTQ+ youth. In some cases, victims are offered in-game currency in exchange for intimate images before being directed to more private online spaces where they are blackmailed.

The sadistic online exploitation (SOE) cycle often evolves as follows:

Targeting a vulnerable user between 8-17 years of age
↓
Building trust or posing as a friend or romantic interest
↓
Obtaining an intimate image or personal information
↓
Threatening exposure to family, friends or school
↓
Demanding further sexual, violent or degrading material
↓
Escalating to animal cruelty, torture or self-harm
↓
Recording and circulating the material within the network for clout

Sextortion is one of the network’s key methods. As noted above, a 764 member may pose as a romantic interest and gradually persuade their target to provide an intimate image or personal information. The victim is then shamed, isolated and threatened with exposure unless further demands are met. Unlike financially motivated sextortion, the objective is generally not money but control, humiliation and escalating violence. Authorities have warned that members are increasingly encouraging victims to commit deadly acts, including suicide, school shootings and other mass-casualty attacks.

Within 764, influence is often tied to the extremity of the abuse a member can make a victim perform. In this sense, violence is also a source of power and influence. The material collected is stored in a ‘lorebook’ which acts as a type of digital currency, and is used to recruit participants and build status. 

An emerging extremist threat

The 764 network forms part of a broader online extremist threat that is gaining increasing attention from child-protection, cybercrime and counterterrorism authorities. On 6 March 2025, the FBI issued a public warning about a sharp increase in activity by 764 and other violent online networks operating in the United States and internationally. By November 2025, the FBI was investigating over 350 people across the United States suspected of involvement with 764 or related networks.

Australia has also made the online extremist threat a national priority. On 29 October 2025, the Australian Federal Police launched Taskforce Pompilid to identify and disrupt online criminal networks targeting young Australians. At the time of its launch, the AFP said it had identified 59 alleged offenders in Australia and that joint operations had resulted in nine arrests overseas and three domestically.

A growing number of international cases demonstrate the range and seriousness of the conduct linked to the network:

In July 2026, Alexis Aldair Chavez was sentenced to 40 years’ imprisonment for racketeering and child-exploitation offences. Chavez was an administrator of 8884, a NVE network related to 764.

In February 2024, a 14-year-old boy in Canada was charged with producing, possessing and distributing child sexual abuse material, distributing intimate images without consent, and making explosives. 

In January 2025, British teenager and 764 member Cameron Finnigan pled guilty to encouraging or assisting suicide, possessing material for terrorist purposes, and possessing child sexual abuse material. Finnigan referred to himself online as ‘Hitler’s child’ and had discussed plots with other members to carry out ‘terror week’.

In April 2025, US authorities arrested Leonidas Varagiannis and Prasan Nepal, who were accused of leading a core subgroup known as 764 Inferno and operating an international child-exploitation enterprise. 

In May 2025, US authorities arrested a 14-year-old boy from Oregon who was accused of planning an explosives attack and mass shooting at a shopping centre in Washington. The FBI said the teenager had shared detailed attack plans online and ABC News reported that he had pledged allegiance to 764.

In October 2025, US prosecutors charged 21-year-old 764 leader Baron Martin with participation in a child-exploitation enterprise as well as conspiracy to kill, kidnap or maim a person overseas. The indictment names nine victims, eight aged between 11-15. Martin was also charged with conspiracy to provide material support to terrorists, the first such charge against an alleged 764 member. 

The diversity of these cases raises questions about whether 764 should be treated primarily as a criminal network, an extremist movement or be formally designated as a terrorist organisation.

To sanction or not to sanction?

In December 2025, Canada became the first country to designate 764 as a terrorist entity, giving its authorities stronger powers to investigate, disrupt and prosecute activities connected with the network. No other country has formally designated 764 as yet, but New Zealand has designated the Order of Nine Angles, an extremist organisation linked to the broader 764 ecosystem.

The treatment of the Terrorgram Collective could serve as a model for other countries considering whether to sanction 764. Terrorgram is a decentralised online network that operates primarily through channels on the messaging platform Telegram. The United States designated Terrorgram and three of its leaders in January 2025, Australia listed it as a terrorist organisation in June 2025, and Canada and New Zealand followed in December. Its designation demonstrates that an online network does not need conventional headquarters or a rigid hierarchy to be treated as a terrorist entity.

Sanctioning 764 could have practical benefits. Depending on the country, designation may criminalise participation or support, enable asset freezes, and grant specialised investigative powers. It would also signal the seriousness and scale of the network’s conduct and warn potential participants that involvement may carry terrorism-related consequences. 

However, Terrorgram presents a more robust case for designation. The organisation espouses a white supremacist and accelerationist ideology and openly provides guidance and instruction for terrorist attacks. By contrast, 764 is more ideologically varied: some participants embrace nihilism, neo-Nazism or accelerationism, while others appear motivated primarily by sexual sadism, cruelty, status or belonging. For these participants, they are not necessarily committing violence to advance an ideological cause, and a terrorist designation might actually be counterproductive, enhancing the notoriety sought by participants and increasing the perceived status of the network.

764 is also more fragmented. It has no stable leadership, formal membership or clear hierarchy, and its participants continually form new groups under different names. Designating 764 may therefore just target a label that can easily be abandoned or replaced without disrupting the underlying network.

Cross-border challenges

Whether or not 764 is formally designated as a terrorist organisation, disrupting it presents a separate practical challenge: its activities routinely cross national and platform boundaries. A victim may be in one country, the offender in another and an intermediary or witness in a third, while the relevant communications and account data are held by technology companies elsewhere.

The investigation into White Tiger demonstrates both the importance and the limitations of international law enforcement cooperation. The German case began after the FBI, while investigating Jay Taylor’s death, traced White Tiger to Hamburg and alerted German authorities. US investigators had identified Shahriar J. as their suspect by mid-2022 and presented their evidence to German police in February 2023. However, he was not arrested until June 2025, more than three years after Jay’s death. 

The reason for the delay was because, according to German law, the FBI material could not be used as evidence in a German court, requiring the investigation to begin again from scratch. This caused considerable frustration among the US investigators, who feared that White Tiger remained free to target further victims. 

The White Tiger investigation shows that identifying an offender is only the first step; the greater challenge may be turning evidence gathered across several countries and platforms into timely action. It also raises a related question: what responsibility should those platforms bear when their services are used to find, isolate and exploit vulnerable children?

Discord’s liability

The Taylor family’s lawsuit focuses on the responsibility of the platform on which Jay was targeted. The case does not merely argue that harmful actors used and communicated on Discord but that Discord’s design, settings, and limited safety measures enabled 764 to locate, isolate and cause harm to vulnerable children.

The complaint states that Discord promotes itself as a place where socially isolated people can find friendship, belonging and community, and permits children as young as 13 to join. The family alleges that, during the relevant period, Discord had no meaningful age or identity-verification system and employed, at most, 79 trust and safety staff. It argues that Discord’s “unmoderated” model, combined with its appeal to vulnerable young users, created conditions that attracted abusers and allowed organised exploitation to develop.

Discord is also designed around private and semi-private interactions, with most servers being small and invitation-only. Although these spaces offer privacy and community, their closed nature can make coordinated abuse harder for parents and platform moderators to detect. Discord does not record voice and video calls, and research suggests that the fleeting nature of real-time communication makes misconduct more difficult to document and investigate.

Jay’s family claims that Discord failed to take reasonable steps to prevent or disrupt child exploitation despite the foreseeable danger to young users. They note that Discord had received numerous warnings of the danger posed by 764 – for example, in July 2021, a 15-year-old girl livestreamed a suicide attempt after being tormented by 764 members. She had reportedly been coerced into carving offenders’ usernames and extremist symbols, including swastikas, into her skin. Jay was targeted approximately six months later.

Government authorities have since raised similar concerns about Discord’s approach to child safety. In May 2026, the Texas Attorney General sued the company, alleging that its product design exposed children to predators while misleading parents and users about the safety of the platform. The lawsuit cites several incidents of violence and abuse, and argues that these were foreseeable consequences of inadequate age verification, unsafe default settings and a largely reactive approach to moderation. Discord disputes these allegations.

The court will ultimately determine Discord’s liability in the case of Jay’s death. But the broader policy question is whether a platform that invites young people to form private communities should be required to design those spaces around foreseeable child-safety risks. The issue is not simply whether Discord should have banned 764 under its rules, but whether it had systems in place capable of detecting and disrupting the network before abuse occurred at all.

The post Sadistic Online Networks Targeting Vulnerable Children first appeared on Online Hate Prevention Institute.

“One of the marks of antisemitism is an ability to believe stories that could not possibly be true” – George Orwell

Executive summary

The 2026 FIFA World Cup hosted old conspiracies in updated garb.

Disputed refereeing decisions, distrust of FIFA, and hostility towards Argentina were assembled into conspiratorial claims that a “Zionist” network had scripted the tournament. Lionel Messi’s visits to Israel and President Javier Milei’s support for Israel were presented as proof of secret influence over FIFA.

The narrative survived contradiction. Spain ultimately won the tournament, yet posts alleging Zionist control continued to circulate. Their reach reveals how major events allow antisemitic ideas to travel through sporting commentary, humour and political solidarity. Event-driven hatred can move faster than systems built around reports submitted after publication.

From sporting grievance to conspiracy

George Orwell described international sport as “war minus the shooting”. Today, international football concentrates national identity, political history and tribal rivalry into a single global spectacle. The World Cup’s scale intensifies that dynamic: every disputed decision can be read as a narrative about power and control.

Distrust in FIFA is well-earned, the peak body mired in decades of corruption. Arguments about Video Assistant Refereeing (VAR) and general refereeing standards, hydration breaks, and political access to referees, bridge pre-existing misgivings into modern political tensions and conspiracy theories.

These dynamics aren’t produced by focal events. They’re already there, waiting for controversy or emotionally charged moments to latch onto. The immediate aftermath of Argentina’s contentious 3–2 victory was like the breaking of a dam. A disallowed goal saw Egypt’s coach Hossam Hassan invoking FIFA’s anti-racism gesture. Supporters experienced the defeat as an injustice with geopolitical relevance, bolstered by centuries-old conspiracy theories about Jewish power and political interference.

Online accounts constructed the defeat as a morality play. Egypt became a proxy for Palestine and Argentina became a proxy for Israel. A football result was described as organised dispossession. “Israel gave them the call”, read one comment that received more than 14,000 likes.

A widely-endorsed comment converts a disputed refereeing decision into Israeli interference.

The conspiracy’s architecture

The story developed through a short chain of association:

·    FIFA is corrupt, so major results may be manipulated.

·    Argentina benefited from contentious decisions, so its progress must have been arranged.

·    Milei supports Israel and Messi has visited Israel, so Argentina becomes “Zionist”.

·    A “Zionist” victory becomes evidence that Jews secretly control FIFA and global institutions.

Each step removes the need for evidence. Public meetings, photographs and political statements become fragments of a hidden system. Fabricated claims then complete the story: Milei was described as Netanyahu’s cousin, a French referee was falsely identified as an Orthodox Jew, and Argentina was accused of transferring Patagonia to Israel.

A “promised 3,000 years ago” joke imports an established antisemitic meme into football discourse.

Four recurring narratives

1. Argentina cheated

Claims that Argentina received favourable fixtures or refereeing decisions were stated as settled fact. Disagreement over individual calls expanded into certainty that the entire competition had been arranged. This provided the bridge from ordinary sporting suspicion to allegations of political control.

2. Argentina represented Israel

Milei’s description of himself as a strongly Zionist president and photographs of Messi in Israel were circulated as evidence that Argentina acted for Israel. Images of Messi with other political leaders received little attention because they disrupted the narrative.

A post describes Argentina and Messi as “Israeli lapdogs” and Milei as a “globalist Jewish weirdo”.

3. Egypt‘s loss as a proxy for injustice and dispossession

Posts cast Egypt as the moral representative of Palestine and Argentina as the agent of a hostile global order. The sporting contest became a battle between humanity and evil. Argentina’s three goals in fifteen minutes disappeared from accounts organised around theft and dispossession.

Football is reframed as a struggle between Palestine and a hostile international system, controlled by shadowy forces.

4. Jewish association proved control

Messi’s visit to the Western Wall, his meeting with Netanyahu and Milei’s support for Israel were edited into montages alleging that Argentina had been “promised” the World Cup. The inference followed a classical antisemitic structure: any relationship with Jews becomes evidence of covert Jewish power.

A post presents Netanyahu as secretly backing Argentina’s World Cup campaign.

Reach and amplification

The examples were distributed through high-reach accounts and recommendation systems. A video asking whether the World Cup was rigged received millions of views. An imam with 825,000 followers drew approximately 2.4 million views for a video addressing whether supporting Messi was religiously permissible because he was allegedly a Zionist. Comments pushing the conspiracy beneath mainstream sports posts received tens of thousands of endorsements.

The comments section performed a central role. Publishers could frame a post around sport or refereeing while users supplied the antisemitic explanation beneath it. High engagement then increased the visibility of the claim and produced the appearance of consensus.

A high-reach religious influencer asks whether support for Messi is permissible because he is allegedly a Zionist.

Why event-driven antisemitism spreads

Major events provide three conditions that favour rapid transmission: intense attention, strong group identities and uncertainty about contested outcomes. Antisemitic explanations arrive as ready-made causal stories. They identify a secret actor, connect unrelated events and convert defeat into deliberate persecution.

The format protects the prejudice. Memes, jokes and sporting insults can carry claims about Jewish control without adopting the tone of ideological propaganda. Users encounter the conspiracy through familiar cultural material, often beneath content produced by outlets and personalities they already follow.

Implications

·    Antisemitic conspiracies gain mainstream reach by attaching themselves to global events.

·    Repeated claims about Jewish control weaken public trust in sport, media and civic institutions.

·    Online abuse can migrate from prominent individuals such as Messi and Milei towards Jewish communities generally.

·    Reactive manual moderation arrives after high-engagement narratives have already shaped public interpretation.

OHPI will publish a longer-form follow-up to this analysis in the coming weeks, examining further examples of this narrative’s spread and setting out potential policy responses for platforms and institutions.

Media contact:
Online Hate Prevention Institute
Email: joshua@ohpi.org.au
Website: www.ohpi.org.au

Join us on social media
OHPI LinkedIn • OHPI Instagram • OHPI Facebook 

The post Antisemitism at the 2026 FIFA World Cup: Argentina Case Study first appeared on Online Hate Prevention Institute.

Content Sensitivity Warning

This briefing discusses the deaths of Aboriginal and Torres Strait Islander people in police custody, including specific cases involving fatal violence. It contains direct examples of racist, dehumanising, and hateful online commentary, reproduced for analytical purposes. 

Readers are advised that names and images of Aboriginal and Torres Strait Islander people who have died are used in this briefing, in line with their portrayal in public reporting and existing media coverage of these cases.

If this content raises concerns for your safety or wellbeing, support is available through 13YARN(13 92 76), a crisis support line for Aboriginal and Torres Strait Islander people, and Lifeline (13 11 14).

As NAIDOC Week comes to a close, OHPI has conducted dedicated analysis focused on the harmful online discourse proliferating around high-profile killings affecting Indigenous Australian communities across the country

In 2019, Kumanjayi Walker was killed by a member of the Northern Territory police, an institution that was later found to bear the hallmarks of institutional racism. A coronial inquest later found the officer held racist views, and was dismissed by police in the midst of a murder trial, from which he was acquitted. There was also the 2022 fatal collision that killed Jai Wright. The case resulted in the first-ever conviction of a NSW Police officer in relation to a death of an Aboriginal or Torres Strait Islander person in custody. Most recently, Kumanjayi White was killed while being apprehended by police in a supermarket in Alice Springs.

Despite their shocking nature, these deaths are nothing new, and are echoes of uncomfortable, unspoken realities about Australia’s past that have been left unaddressed and are now spilling out into the digital realm. This hatred and toxicity has marred Australia since the arrival of the First Fleet and the brutal treatment of Aboriginal and Torres Strait Islander peoples that followed.

The circumstances surrounding each killing discussed in this briefing are unique, and the facts of each case are worth exploring in detail in their own right. Rather than re-litigating the specific facts of these preventable tragedies, this briefing instead explores the drivers and dynamics behind the charged online discourse that surrounds them. In drawing attention to these cases, we believe important lessons can be drawn about how harmful narratives (at times spurred by traditional media) are further exacerbated by the dynamics of the digital arena, where abusive content and their toxic comment sections thrive, and where mis/disinformation spread freely. It is this poorly governed digital public square that makes platforms feel unsafe for users, and which is detrimental to wider society.

The cases above all follow a similar pattern, one where a death at the hands of the state becomes the flashpoint for a recurring cycle of public reaction. This cycle typically unfolds as follows:

1. Avoidable tragedy or atrocity
2. Community outrage
3. Backlash to that outrage (online hate)
   1. steps 2 & 3 repeat in a continuous, self-reinforcing feedback loop, resulting in further toxic online discourse
4. Institutional attempts at accountability and justice
5. Where those institutional steps fail, further social backlash and instability. 

Note on Step 3: This step is sometimes a precursor to 1 and 2, and frequently recurs throughout steps 4 and 5 as well. Online hatred is near-inevitable in the aftermath of atrocities of this nature. As seen with the Bondi Beach attack, online spaces can flood with hatred towards the victimised community almost immediately, often before that community has had any opportunity to express its own grief or outrage. 

These incidents, particularly once they go viral, leave long lasting ripple effects that combine in the digital space to shape the wider landscape of online discourse. While institutional accountability can take years to materialise (Kumanjayi Walker’s coronial inquest findings were released in June 2025, 6 years after his death), these cases fester in the online zeitgeist throughout, fuelling shock-jocks and grifters who monetise culture-war content. 

The ‘digital’ and the ‘real’ world are now one and the same

Online platforms are taking steps away from automated moderation of online hate speech, increasingly leaving the reporting and moderation of harmful online discourse to individuals. In practice, this places the burden onto the people most likely to be the victims of that harmful speech, who are now responsible for reporting it themselves, with these reports often rejected and the harmful content remaining online. Where platforms do rely on AI for content moderation, significant questions remain about its safe implementation.

Research by the AHRC notes that the online world “creates a space of hypervisibility in which the ‘real’ and ‘virtual’ worlds overlap”. In this environment, the distinction between digital hate and real-world consequence is blurred, and hate seeps freely into people’s day-to-day lives.

A survey conducted by Macquarie University showed that 88% of Indigenous people witnessed examples of racism online. Much attention has already been paid to the marked rise in racist online rhetoric in the wake of the failed 2023 Voice referendum. High-profile Indigenous Australians are exposed constantly to racist abuse. Spend any time online and comment sections can be seen to devolve into spaces of unchallenged hostility toward entire groups. Indigenous public figures are subjected to degrading and racist material on a daily basis, posted by faceless accounts that evade accountability behind the veil of online anonymity.

The risks of this ungoverned ‘information superhighway’ are clear: without sensible governance, we are sleepwalking towards disaster. 

Defenders of the social media platforms and the gatekeepers of the online information space point out that they are not the ‘arbiters of truth’, and have no interest in limiting free speech on the basis of political opinion. But when the spaces these platforms control reinforce racist perceptions, the consequence can result in real-world violence.

Addressing these wicked problems facing our First Nations communities requires multi-pronged policy response across several domains, including areas like police accountability, media standards, and a general national commitment to truth telling. The scope of this briefing, however, is narrower. It examines the patterns identified above (particularly on X and Meta), whereby the character of Indigenous victims of violence is put on trial in the court of public opinion, alongside the broader stereotyping, victim-blaming, and hostile commentary that justifies or minimises these atrocities.

It is within this context that this briefing explores a central question: what do social media platforms owe their users? Unimpeded access to a digital space with no limits on what a user encounters? Or a restricted internet, where platforms censor content according to the preferences of individuals and governments?

This is a false binary. Between unrestricted access and heavy-handed censorship lies a middle ground where sensible governance has a role to play – namely, a digital duty of care.

Note: Some examples below do not meet the threshold for hate speech, and are instead illustrative of awful-but-lawful or ignorant content. Still, they reflect the deluge of online abuse that platforms have facilitated in the wake of these events.

Case Study 1 Kumanjayi Walker 

Zachary Rolfe killed Walker at point blank range after Walker stabbed him in the shoulder with a pair of scissors during a failed arrest. Rolfe was later acquitted on the grounds of self-defence. However, Rolfe’s inquest revealed the extent of his pre-existing racist attitudes and beliefs. He shared illegally obtained bodycam footage of physical altercations with Indigenous Australians, distributing it to family and friends as something to be proud of. Rolfe and the subculture that could develop at NT Police are a product of this heightened hateful environment, an embodiment of our collective failure to adequately protect Australia’s most vulnerable minority groups.

Mainstream coverage of the trial that led to Rolfe’s acquittal portrayed Walker as little more than an habitual criminal. The coverage was widely criticised, and described by the political editor at The West Australian as “unethical, insensitive, victim-blaming, irresponsible” and “a national disgrace”.

Digital coverage and online responses were no more forgiving. OHPI’s research has shown that events of this type correlate with distinct spikes in online hate. We can clearly see how the platform mechanics turn social media into a hate production line, exposing entire populations to hate at scale. 

SNAICC – National Voice for our Children, the national peak body for Aboriginal and Torres Strait Islander children (and a regular recipient of online hate on its posts) notes that Meta’s moderation tools are “often insufficient to address the scale and evolving nature of this issue.” Just with other types of online hate, commenters try avoid automatic filtering (via keywords) by deferring to innuendo, coded language or indirect hostility or the under guise of ‘political speech’ (figleafing). SNAICC’s submission to the Parliamentary Inquiry Into Racism, Hate and Violence Directied at Aboriginal and Torres Strait Islander People also contains, from page 37, a snippet of the deluge of hateful comments. 

As mentioned above, problematic online discourse surrounding the killing of Kumanjayi Walker involved stereotyping, victim-blaming, and hostile commentary that justified or minimised the killing. 

Example Post 1

Context: A post on X framing the deceased as ‘the boy nobody wanted’, claiming that race played no role in his death, contrary to the facts uncovered during his inquest. The account is Australia-based and posts prolifically: at the time of writing, the account has posted 21 tweets in the last 24 hours, about various hot button culture war issues. The divisive, AI slop-laden posts and reposts make sweeping generalisations about large groups of people, and decries the oncoming ‘complete societal breakdown in the UK’. The tweets also contain reposts of diatribes about the ‘monoculture’, and responses like the following.

Example Post 1 Comments

One response to the above tweet stereotypes Indigenous Australians as alcoholics, a trope historically used to dismiss and minimise Indigenous disadvantage, trauma, and the structural harms. By reducing a complex, historically rooted issue to an individual moral failing, the post trivialises the real and ongoing harm faced by Indigenous communities.

Comments from troll accounts, justifying the killing of an individual without any consideration of proportionality or justice: 

An additional comment that praises the actions of Rolfe: 

Case Study 2 Kumanjayi White

Northern Territory police killed 24-year-old Kumanjayi White, a Warlpiri man with cognitive disabilities, during a confrontation in a Coles supermarket. White’s death sparked outrage in Alice Springs, accompanied with a litany of online hatred directed towards White and the wider Indigenous community. 

Example Post 2

This news post from 26 May from the National Indigenous Times attracted comments that delivered ad hominem attacks against White, suggesting he was deserving of his fate. 

Example Post 2 Comments

The ad hominems below, against the victim of an extrajudicial police killing, are reminiscent of the online discourses proliferating after the killing of George Floyd. 

Example Post 2 Comments (cont.)

The next comments echo the same fallacious logic seen in analogous cases overseas, where minorities have been killed while committing trivial crimes. The underlying assumption is that because the victim was allegedly committing a petty crime, proportionality becomes irrelevant, justice has been served, and those responsible for an entirely avoidable extrajudicial killing are exonerated.  

Case Study 3: Jai Wright

A NSW Police officer was convicted over the death of Jai Kalani Wright, a 16-year-old boy killed when his motorbike collided with the unmarked police car. The case represents the first time a NSW Police officer was convicted in relation to a death in custody of an Aboriginal or Torres Strait Islander person.

Example Post 3

This post about the conviction by digital newspaper the National Herald attracted 464 comments. The overwhelming majority are critical of Wright for having “committed a crime”, critical of “Indigenous parenting” (e.g. ‘teach your children to stop for police’), or critical of the reporting itself. The latter example shifts blame onto Wright for his own death, constituting victim-blaming, and it invokes the oft-repeated claim of “fake news” for information that doesn’t align with the commenter’s worldview – which amounts to atrocity denial or minimisation). 

Example Post 3 Comments

Example Post 4 (with Comments)

This reply to a post regarding Wright involves incitement to contempt and endorsement of harm towards First Nations people.  

This response attributes individual conduct to the racial group as a whole, and minimises the validity of the real suffering faced by Indigenous Australians. 

Potential ways forward in Digital Governance

The hope that accountability and positive social change are inevitable for preventable First Nations deaths in custody is undermined by wider social and technological dynamics, particularly around digital governance. While this ongoing inquiry continues, it has been 35 years since the Royal Commission into Aboriginal Deaths in Custody, and since then, at least 630 people have died in custody.

The submissions that have been tendered to the more recent inquiry give insight into the severity and wickedness of the problem, as well as various informed perspectives and policy proposals to target the problem of online hatred. Some of these proposals could help in the combatting of all online hate, not just that directed towards First Nations communities. 

The AHRC have recommended the introduction of a digital duty of care, requiring social media companies to “identify, assess and mitigate foreseeable risks arising from recommender systems and monetisation practices that incentivise the amplification and normalisation of racist narratives”.The submission by Amy Rust of Jindawayni Consulting echoes this sentiment, suggesting the introduction of ‘duty of care obligations’ for organisations that host public comment platforms that facilitate racial abuse. In her words, ‘Freedom of expression does not require providing a platform for racial abuse’. 

The National Anti Racism Framework of 2024 recommends that the government introduce “effective legal protections against online hate, with particular attention given to regulation and enforcement against, and liability of, digital platforms” to be developed in consultation with First Nations and other “negatively racialised communities”. This could be achieved through the establishment of a “National Indigenous Digital Safety Taskforce”, as suggested by the Australian Multicultural Action Network’s submission. This calls for a mechanism through which governments, digital platforms, indigenous organisations and civil society groups collaborate on combatting online racism towards First Nations peoples. 

ASIO’s submission is frank in its assessment that “The increasing normalisation of, and access to, extremist content online in combination with high levels of intimidating, violent and intolerant rhetoric increases the exposure of extremist ideas and likely serves as a gateway for some individuals to violent extremist beliefs and ideals”… which could inspire individuals to commit violence.

Racism may have played a part in Walker’s death, but it doesn’t have to play a part in his legacy; the ongoing abuse faced by First Nations peoples after these senseless killings must be addressed. Before we can begin to have an honest and constructive national conversation and truth telling about the issues plaguing our First Nations people, we need to fix our feeds and our online public square, and put out the fires that are stoking real world violence. 

If this content raises concerns for your safety or wellbeing, support is available through13YARN(13 92 76), a crisis support line for Aboriginal and Torres Strait Islander people, and Lifeline (13 11 14).

Media contact:
Online Hate Prevention Institute
Email: joshua@ohpi.org.au
Website: www.ohpi.org.au

Join us on social media
OHPI LinkedIn • OHPI Instagram • OHPI Facebook 

The post NAIDOC WEEK IN FOCUS – The Online Hate Spiral Targeting Aboriginal and Torres Strait Islander Peoples first appeared on Online Hate Prevention Institute.