A HEIA Thematic Report / Report # 2
Primary Author: Jenni Long
Download this report in PDF format:
Executive Summary
This report explores the relationship between hate speech and Harmful Inaccurate Information (HII) in Aotearoa’s online spaces to determine if and how they modify one another.
We consider overlaps within a database of 2183 New Zealand-based hate speech posts and 3436 New Zealand-based HII posts collected from Facebook, Reddit, Instagram and Telegram through 26 November 2023 - 11 March 2024.
Our methodology includes a two-step verification process, first utilising two Large Language Models (LLM) to detect HII and hate speech posts, which were then manually verified and categorised.
We demonstrate that 1) antisemitic hate speech contains the greatest amount of HII compared to the other categories in our dataset, 2) social groups are used as scapegoats within HII, 3) HII can converge different strands of hate speech, and 4) where HII and hate speech combine, population/cultural replacement is a recurrent theme.
Contents
Harmful Inaccurate Information (HII)
Harmful, Inaccurate Information (HII)
Analysis of HII and Hate Speech relationship
2. Anti-LGBT+ Hate Speech and HII
4. Anti-Māori Hate Speech and HII
5. Anti-immigrant Hate Speech and HII
7. Anti-Asian/Southeast Asian Hate Speech and HII
8. Anti-Black Hate Speech and HII
9. Ableist Hate Speech and HII
Key findings and implications for Aotearoa
Antisemitic hate speech overlap with HII
Social groups as scapegoats in HII
HII congealing multiple enemies (and alliances)
Table of Figures
Figure 1. Number of hate speech posts by category
Figure 2. Number of hate speech posts over time by category
Figure 3. Breakdown of antisemitic hate speech which includes other forms of hate speech
Figure 4. Breakdown of anti-immigrant hate speech which includes other forms of hate speech
Figure 7. Number of hate speech posts compared to number of HII posts over time
Figure 8. Number of hate speech posts also containing HII over time
Figure 9. Number of anti-LGBT+ hate speech posts that also contained HII over time
Figure 10. Number of antisemitic hate speech posts that also contained HII over time
Figure 11. Number of anti-immigrant hate speech posts that also contained HII over time
Whakataki – Introduction
Hate and Extremism Insights Aotearoa (HEIA) conducts data-led research to measure and analyse harmful online rhetoric. HEIA is based at the University of Auckland and led by Dr Chris Wilson. This report analyses the relationship between hate speech and Harmful Inaccurate Information (HII) in New Zealand, utilising online data collected from late November 2023 through to March 2024.
Harmful Inaccurate Information (HII) is false ideas, beliefs and claims spread in New Zealand which are capable of causing a range of harms to individuals, communities and institutions. Hate speech describes the vilification, abuse, and/or dehumanisation of an individual or community on the basis of their culture, ethnicity, or other social group membership.
Both HII and hate speech can generate polarisation, mistrust, degrade social cohesion and lead to violent extremism. The two often overlap, and when they do so this may increase the harm they cause (Gartenstein-Ross et al., 2023). The specific links between online HII and hate speech, however, are under-researched.
Both HII and hate speech can generate polarisation, mistrust, degrade social cohesion and lead to violent extremism.
This thematic report employs large data techniques to generate a comprehensive snapshot of HII and hate speech in Aotearoa over a 4-month collection period. We use this period to consider the volume and content of HII and hate speech, examining how and when they overlap.
This report will begin with a background section to contextualise our findings and show why this research is important for mitigating harm. After clarifying our methods, we then look deeply into the quantitative and qualitative features of our dataset, making sense of specific spikes, themes, and patterns in posting HII and hate speech. Finally, we describe our key findings derived from this analysis, which in turn, will allow for more targeted and effective interventions for preventing and countering harm in Aotearoa.
We have included a small number of anonymous posts which represent the HII and hate speech themes discussed. While these by no means represent the most offensive messages in our dataset, and we have redacted offensive terms, readers should exercise caution before reading. |
Hate Speech HEIA defines hate speech as speech that goes beyond highlighting differences in practices and identities between groups, and vilifies, demeans, abuses or dehumanises people on the basis of their gender, ethnicity, religion, disability status or other identity. In doing so, hate speech is therefore different to political opinion and debate. |
Harmful Inaccurate Information ‘Harmful Inaccurate Information’ (HII) refers to all misleading information which causes harm or which is likely to have a harmful impact, regardless of whether the spreader of that information is a) aware of its misleading nature or b) intending to cause harm. |
Not all hate speech is HII While hate speech involves slurs and abuse, it can occur without including the spreading of inaccurate information as fact. In such cases, this speech is abuse or opinion (as heinous and ignorant as others may find it) and not HII. Example: ‘Yes, [homophobic slur should] be smacked over starting with you’ (21/12/2023). |
Not all HII is hate speech A lot of HII contains false and misleading claims which are harmful in a range of ways but which do not demean a particular community. These statements are HII but not hate speech. Example: ‘Mark of the beast is upon us covid-19 real meaning is certification of vaccinations by ID identification. (-19) means it's a patented Bioweapon not a Vaccine’ (27/02/2024). |
In some cases HII and hate speech overlap In these cases, the speaker demeans a particular group and spreads inaccurate information. Example: “[x group] control the mainstream media, social-media, banking systems and are in key positions of political power in all White societies. It is very obvious that they are not ‘victims’ - they are a hostile aggressive tribe with an innate biological drive to dominate and subjugate” (08/02/2024). |
Background
Hate Speech
Hate speech targets social groups, with the consequence of isolating and threatening the safety of target communities (Udupa et al., 2020). Hate speech goes beyond highlighting differences in practices and identities between groups, and dehumanises people on the basis of their gender, ethnicity, religion or other identity. Simultaneously, hate speech reinforces a sense of sameness within the ‘in-group’, thereby increasing polarisation (Castaño-Pulgarín et al., 2021).
Perpetrators can justify the use of hateful rhetoric by minimising the harm caused, assigning blame to the victim (Faulkner & Bliuc, 2016). Hate speech is different from political debate because it villainises people within a target group on the basis of their membership in that group (Castaño-Pulgarín et al., 2021). This report will consider the most insidious examples of hate speech, which vilify, abuse, or dehumanise targets.
Observers often claim that ‘hate speech is on the rise’, but this is typically difficult to prove. The amount of online content is vast, and so accurately measuring hate speech is near impossible – particularly over long periods of time (Udupa et al., 2020). This is further complicated when the definitional boundaries between hate speech and political or antagonistic debate are controversial and blurred (Udupa et al., 2020). However, the proliferation of new information technologies facilitates greater dissemination of hate through features such as online information echo-chambers (Johnson et al., 2019; Milhazes-Cunha & Oliveira, 2023).
Because the overall level of hate speech within online spaces is difficult to measure, it is also hard to make evidence-based claims about the role of hate speech for increasing hate crime. This is particularly true in Aotearoa. New Zealand Police have only recently established a register which records the number of hate crimes occurring, so knowledge surrounding the victims and perpetrators are limited (Shand, 2021; Xia, 2024). Consequentially, we have little information regarding the rise and fall of hate-based violence in response to offline events, let alone how online hate speech influences the degree and form of hate crime occurring in Aotearoa.
Even so, the normalisation of hateful beliefs is one precursor to hate-based violence, and online hate speech has been linked to extreme real-world acts of violence (Leitch & Pickering, 2022; Williams et al., 2019). For example, the perpetrator of the 2019 Christchurch Mosque attacks had spent at least five years immersed in Islamophobic, antisemitic and racist online discourse (Wilson et al., 2024). The relationship between online hate speech and offline violence is bidirectional and transnational (Williams et al., 2019). A study observed a 250% increase in racial hate speech following the death of George Floyd, and during the same time period, race-based hate crimes in the United Kingdom rose by 6% (Flora & Jemma, 2020; Lupu et al., 2023). Tracking the targets of hate speech is therefore important for predicting potential victims of offline violence.
The normalisation of hateful beliefs is one precursor to hate-based violence, and online hate speech has been linked to extreme real-world acts of violence
Harmful Inaccurate Information (HII)
Misinformation is false content spread without any intention to mislead or cause harm. Disinformation is the dissemination of information with the knowledge that it is false and likely to cause harm (Gibbons & Carson, 2022; HEIA, 2023; Wang et al., 2022).
Because it is often difficult to know the intent of those who spread false ideas, discriminating between misinformation and disinformation online is challenging. Hereby, in this report (as in our previous reports) we use the term ‘Harmful Inaccurate Information’ (HII) to refer to all misleading information which causes harm or which is likely to have a harmful impact, regardless of whether the spreader of that information is a) aware of its misleading nature or b) intending to cause harm. This broad definition is outcome focused, and is best suited to data collection for an analysis of the problem of HII in Aotearoa’s online spaces (HEIA, 2024). For an in-depth definition of HII, see ‘Harmful Inaccurate Information: Defining disinformation, misinformation’.
It is difficult to discuss these topics without considering freedom of expression, which is essential for fair, democratic societies. Free speech encourages critical thinking, encourages a diversity of viewpoints, and enables people to publicly express their ideas in order to gain feedback and foster connection with others (Herz & Molnar, 2012). Freedom of expression is an essential value for supporting civil society and liberal democracy, but so too is protecting vulnerable communities and individuals from identity-driven harm, and defending their capacity to fully participate in society.
HEIA does not recommend the regulation of hate speech and HII. Suppression of speech through regulation and content removal can undermine rights of free expression, risks disproportionate targeting of certain (legal) political and other opinions, and causes backlash and grievance, acting to solidify divisions and inflame feelings of persecution, undermining trust. This may only prompt the spread of further hateful rhetoric (Udupa et al., 2020). Furthermore, this approach can encourage the movement of those with alternative opinions to less mainstream platforms where more insidious perspectives are rife (Johnson et al., 2019).
However, HEIA asserts that it is wise to track and monitor narratives of hate and HII within Aotearoa’s online spaces (while maintaining poster anonymity), to establish baselines and highlight notable trends and patterns. This helps prepare us for future increases in damaging inaccurate information, providing agencies and others evidence to work with rather than assumption.
Hate Speech and HII
Hateful content is more common in times of crisis and social unrest, and often spikes in response to related offline events (Lupu et al., 2023). Categories of hate speech unrelated to the offline event can also spike, indicating a general response driven by uncertainty and anxiety. This is significant, because anxiety is also a core cause of HII, which tends to arise in uncertain contexts or unforeseen circumstances. In this way, online hate speech and HII often rise in tandem (Udupa et al., 2020). The spread of HII causes confusion and anxiety in itself, creating a disorienting environment which is fertile ground for hate towards marginalised groups (EFSAS, 2021).
The spread of HII causes confusion and anxiety in itself, creating a disorienting environment which is fertile ground for hate towards marginalised groups
Fact-checking services which determine the credibility of claims made on social media websites observe a strong overlap between HII and hate speech (Hameleers et al., 2022). Claims which contain more false information are also more likely to contain hateful rhetoric, thereby indicating that HII and hate speech may amplify one another (Hameleers et al., 2022). This is likely due to ‘scapegoating’, where a person or group is unfairly blamed for a problem or event (AEC, 2023).
Both hate speech and HII play a role in polarising societies, widening the gap between social groups and making dialogue more difficult (Arce-García & Menéndez-Menéndez, 2023; Cinelli et al., 2021). Alongside their ability to generate polarisation, HII and hate can bring together previously antagonistic groups united in common cause and enemy, as observed in the protests at the Beehive grounds inspired by its Canadian counterparts (Robie, 2022; Vasist et al., 2023). The globalisation of HII can facilitate the importation of ideas which justify hate towards local diaspora communities (Cosentino, 2020). For example, alongside the uptake of globalised conspiracy theories which blame Chinese culture for the Covid-19 pandemic, Aotearoa has experienced a surge in anti-Asian hate.
HII can also be employed intentionally by individuals and groups seeking to undermine groups and institutions. Marwick and Lewis (2017) show that in online spaces, there are increasing links between groups spreading misogynistic, Islamophobic and antisemitic hate and conspiracy theorists.
The spread of combinations of HII and hate can have violent implications. The perpetrator of the Charleston church shooting was influenced by untrue, and exaggerated information about black on white crime (Daniel, 2018). Far-right HII fuelled the Charlottesville rally and consequent violence targeting a counterprotest (Daniel, 2018). It was also crucial in the ‘stop the steal’ movement and subsequent attack on the Capitol. In India the spread of rumours online has led to mob lynchings of Muslim minorities by Hindutva nationalists (Mirchandani, 2018).
HII and hate speech are therefore closely linked, following similar patterns and are caused by similar phenomenon. What is less clear are the ways in which hate speech and HII interact within online spaces and the consequences of these interactions - this gap is particularly present in Aotearoa. These uncertainties inspired the aims of this report, which are to 1) identify where HII overlaps with hate speech in Aotearoa’s online spaces; 2) clarify the products of HII and hate speech interactions; and 3) to determine if, and how, HII generates relationships between different forms of hate.
Methodological notes
This report uses two proprietary Large Language Models (LLM) to detect HII and hate speech within a collection of 1,546,742 posts from online social media platforms including Facebook, Reddit, Instagram and Telegram.
The hate speech LLM identified posts based on the input: ‘Does this poster show hatred towards, or vilify, demean, abuse or dehumanise an individual or community because of their cultural or ethnic identity or disability status or physical appearance?’. The HII LLM identified posts based on the input: ‘This post must include inaccurate or false information that will deceive people and is likely to cause harm. This includes conspiracy theories, which are false, incorrect and untrue claims of a powerful group of people enacting a plan to increase their own power which has damaging consequences for the majority of people or society’.
The posts identified as HII and/or hate speech by the LLMs were subsequently manually verified by a team of four researchers who removed those which did not meet our high threshold of inaccuracy and harm. The contents of the resulting posts were analysed and categorised based on specific key themes and subthemes alongside previous literature on hate and HII. Regular meetings and comparison ensured coding was consistent between team members.
All data are anonymised on collection and no personal information is intentionally collected or used in HEIA’s research. Any personal information inadvertently collected is deleted and we do not provide any such information or other data to third parties nor use it in our analysis.
We draw attention to the fact that the numbers of hate speech and HII within our dataset are low. We do not suggest that this is representative of the level of hate faced by target communities in Aotearoa. We collect from public areas on particular platforms only - some of the most extreme hateful sentiments are posted in private (rather than public) online spaces. In addition, our methodology is rigorous, with strict benchmarks for ‘harm’ and ‘hate’ alongside a two-step verification process for weeding out posts which don’t meet these benchmarks (ensuring those simply expressing a political or other opinion are excluded for example), resulting in a database that consists of the most important HII and hate speech. These low numbers are manageable, allowing us to conduct a more targeted analysis of the state of HII and hate speech in Aotearoa. Our aim is to identify observable patterns and trends related to the HII/hate speech relationship, rather than provide a complete picture of the levels of hate speech and HII.
Over the collection period, 19,058 posts were identified (by LLMs) as containing hate speech, and 26,297 posts identified as containing HII. Once the resulting posts were manually verified, our database consisted of 2183 New Zealand-based hate speech and 3436 New Zealand-based HII posts.
Some posts were marked as both HII and hate speech. HII is present within: 5.3% of misogynistic hate speech, 16.5% of anti-LGBT+ hate speech, 27.5% of antisemitic hate speech, 8.3% of Islamophobic hate speech, 7.3% of anti-Māori hate speech, 7.69% of anti-Asian/southeast Asian hate speech, 8.5% of white supremacist/anti-black hate speech, 1.08% of ableist hate speech, and 6.2% of anti-immigrant hate speech. An analysis of these overlaps is below.
Within this report we include examples of posts from our dataset which contain HII and hate speech. We have carefully considered the merits and problems with reproducing posts, but believe their inclusion within our discussion is important to adequately demonstrate the vitriol directed towards target communities and its relationship to HII. Some posts have been counted in both datasets.
Quantitative results
Harmful, Inaccurate Information (HII)
After manually verifying the 26,297 posts identified by the LLM for the reporting period of 26 November 2023 - 11 March 2024, our total dataset is 3436 New Zealand-based HII posts.
HEIA identifies four overarching categories of HII: 1) Health and Covid-19, related to distrust in the official health narratives and initiatives of the New Zealand Government and international health organisations; 2) Globalist Conspiracies, which relates to the belief that world events, including in New Zealand, are being controlled by an elite coalition of ‘globalists’ intent on creating a ‘New World Order’; 3) Politics (Distrust and Withdrawal), which maintains the New Zealand Government is undemocratic and/or cannot lawfully control individuals; and 4) Social Anxiety, which is associated with a belief that society is moving in the wrong direction.
Each of these themes contains several more specific subcategories. For a full explanation of how HEIA categorises HII data, see the see ‘Trends in Harmful Inaccurate Information in New Zealand’.
Hate Speech
After manually verifying the 19,058 posts identified by the LLM for the reporting period of 26 November 2023 - 11 March 2024, our total dataset comprises 2183 New Zealand-based hate speech posts.
Within this total, the following forms of hate speech can be observed: misogyny (186 posts), anti-LGBT+(513 posts), antisemitism (577 posts), islamophobia (72 posts), anti-Māori (478 posts), anti-Asian/southeast Asian (143 posts), anti-black and/or white supremacy (82 posts), ableist (92 posts), and anti-immigrant (239 posts) (see fig. 1).
Figure 1. Number of hate speech posts by category
We also identified 98 posts which were directed towards other out-groups not specified within our above categories. For example, a few posts specifically targeted Native Americans, and others targeted Pasifika peoples.
Overall, verified hate speech represents a very low volume of HEIA’s collection of New Zealand-based posting – about 0.15%. Within our collection period, we observed a decline in the number of hate speech posts through mid-late December 2023 (Fig. 2). This is likely because December is holiday season during which people spend less time online.
From 4/1/2024 however, overall hate speech numbers rose, with peaks on specific days such as 8/1/2024, 12/1/2024, and 16/1/2024, and a sharp increase on 19/1/2024. This day in particular saw significant numbers of anti-Māori hate speech surrounding the disestablishment of the Māori Health Authority. Similarly, a spike in antisemitic hate speech occurred on February 27th-28th in response to the death of a prominent Jewish investment banker.
Figure 2. Number of hate speech posts over time by category
Figure 3. Breakdown of antisemitic hate speech which includes other forms of hate speech
A number of posts were marked as multiple categories.
21.44% of anti-LGBT+ hate speech contained another category of hate speech, with the most significant overlap with misogynistic hate speech, followed by antisemitism.
15.4% of anti-Māori hate speech contained another category of hate speech, where misogynistic hate speech represented the most significant overlap followed by anti-LGBT+ and antisemitism.
12.8% of antisemitic hate speech contained another category of hate speech, and anti-LGBT+ hate speech represents almost half of this overlap (see fig. 3).
40.5% of anti-immigrant hate speech contained another category of hate speech, with significant overlap with anti-Asian/southeast Asian hate speech, Islamophobic hate speech, and anti-Black hate speech (Fig. 4).
Anti-immigrant hate speech sees a surprisingly high overlap with misogynistic hate speech, which was mostly observed following shoplifting charges against a former female MP in January 2024. This event saw a sharp increase in both anti-immigrant and misogynistic hate speech, as well as convergence between these two forms (See 17/01/2024 in fig. 5). Similarly, a spike in combined antisemitic and anti-LGBT+ hate speech occurred on 9/12/2023, when a series of posts were shared which targeted Jewish doctors involved in gender affirming care and/or surgery (See fig. 6).
Figure 4. Breakdown of anti-immigrant hate speech which includes other forms of hate speech
Figure 5. Anti-immigrant hate speech over time vs anti-immigrant hate speech posts also containing other forms of hate speech
Figure 6. Antisemitic hate speech over time vs antisemitic hate speech posts also containing other forms of hate speech
Analysis of HII and Hate Speech relationship
Figure 7. Number of hate speech posts compared to number of HII posts over time
Like hate speech, the amount of HII online decreased through mid-late December (see fig. 7). It does, however, come in much sharper spikes than with hate speech. The period which saw the greatest increase in both HII and hate speech - early February to early March 2024 - also saw the greatest number of posts which contained overlaps between them (see fig. 8).
"The period which saw the greatest increase in both HII and hate speech [was] early February to early March 2024"
Figure 8. Number of hate speech posts also containing HII over time
1. Misogyny and HII
Misogynistic hate speech targets women as a social group. Of the total 2183 posts that we identified as containing hate speech, 186 were categorised as misogynistic. This number is relatively low compared to other categories, likely because misogyny generates hostility towards specific types of women – rather than all women - who act in ways which do not align with traditional expectations of femininity (Manne, 2017). Hate speech is one of many manifestations of online gender based violence against women (OGBV), but other forms include online harassment, sexualisation and objectification, impersonation, stalking, trolling, and doxing (Bundtzen, 2023).
10 of the total 186 misogynistic hate speech posts within our dataset also contained HII sentiments. Within these posts, a common theme is the assertion that Aotearoa society favours women over men. Representative posts claim that women can perform poorly and be successful, and men who complain face difficulties: “Females and **** walk around with the clip boards and earn the bucks even if they **** up. In NZ, the receptionist earns(s) more than the tradesmen” (19/02/2024). This notion is connected to ideas disseminated in the transnational ‘manosphere’ - the diverse yet interconnected online network of anti-feminist, misogynistic websites (Ging, 2017, p. 638). These perspectives have found footing in Aotearoa. This can be observed within our dataset where misogynistic hate speech combines with HII to misconstrue the intentions of women to gain control over men.
Whilst misogynistic hate speech was observed in relatively low numbers in comparison to other forms of hate speech, 432 posts within our total dataset contained more general misogynistic sentiments, which disparaged specific individuals through the lens of gender, rather than women as a social group.
21 posts which displayed misogynistic hate speech or general misogynistic sentiments also included HII. 12 of the posts contained social anxiety themes, generating connections between misogyny and anti-LGBT+ hate speech. For further analysis of this relationship, see HEIA’s report on ‘The Relationship Between Misogyny, Anti-LGBT+ Hate Speech, and Harmful Inaccurate Information in Aotearoa’.
2. Anti-LGBT+ Hate Speech and HII
Anti-LGBT+ hate speech refers to rhetoric which targets people of diverse genders or sexualities as a social group, and represented the second highest proportion of hate speech within our dataset. Of the total 2183 posts that we identified as containing hate speech, 513 contained anti-LGBT+ hate speech. 85 of these also contained HII sentiments.
81 of these were posted within a period of one month through 11 February to 11 March 2024 (see fig. 9). The subject matter being discussed within this period was wide-ranging, from the role of transgender people within competitive sports overseas to sex education within Aotearoa. One frequently discussed topic, however, was a claim made in mid-February by University of Sussex Hospitals NHS Trust that breastmilk from transgender women is comparable to cisgender women (Searles, 2024). This event prompted an increase in anti-LGBT+ HII within our data, referring to child abuse and other topics.
Figure 9. Number of anti-LGBT+ hate speech posts that also contained HII over time
68 of the posts contained social anxiety themes surrounding perceptions of a widespread gender ideology being implemented within schools, governments, and international institutions, reminiscent of cultural warfare ideas imported from the United States: “The LGBTQIA never ending+ indoctrination and **** agenda came from the Universities with cultural Marxism and postmodernism ideology that is infecting and destroying societies and countries from within” (14/02/2024).
Another common trend is the denial that transgender people exist, and they are instead a product of either poor mental health or fetish. At its most insidious, this intersection of HII and hate suggests that transgender people – or LGBT+ communities in general – are involved in the abuse of children.
35 posts marked as anti-LGBT+ hate speech also contained globalist conspiracies, suggesting that the so-called ‘globalist elite’ are transgender and/or wish to implement a gender ideology as a means of world domination: “[…] they're removing basic biology from our schools as it interferes with the **** narrative, and now they're coming for the children to indoctrinate them, this will end up turning children against their parents” (26/02/2024)
In line with this conspiracy, there is overlap with antisemitism, where 9 posts also frame Jewish people as being involved in this elite group: “Amusing, isn't it, how all this LGBQ+, Gender Reassignment **** Agenda matches (((their)))[1] "teachings" and "beliefs". Just another “Coincidence". Nothing to see here, *****. Move on” (04/02/2024). Triple brackets are used to refer to members of the Jewish community.
3. Anti-Semitism and HII
Antisemitic hate speech targets Jewish people as a social group, and represented the highest proportion of hate speech contained within our dataset. Of the total 2183 posts that we identified as containing hate speech, 277 contained antisemitic hate speech posts. Of these, 159 also contained HII sentiments, representing the most significant overlap between HII and hate speech within our study. Antisemitism often integrates conspiracy theory, depicting Jewish people as “’crafty, greedy, selfish outsiders’ who are conspiring against us […] to take what is ours, to control our lives and futures, to ‘replace’ us” (Finkelstein et al., 2020, p. 2)
106 of these were posted between 11th February 2024 to 1st March 2024 (see fig. 10) with a significant spike between February 26th-29th in response to the death of a prominent Jewish investment banker. The wealthy family of this prominent investment banker has played a central role in antisemitic conspiracies since the nineteenth century, representing “a symbol of the rising power of Jews under modern capitalism”(Langer, 2022, p. 24).
[1] In antisemitic online spaces, putting three brackets around a word is a head nod to Jewish people
Figure 10. Number of antisemitic hate speech posts that also contained HII over time
Whilst this spike was related to an offline event, antisemitic hate speech spikes even without relevant offline events to prompt them. The bulk of antisemitic hate speech is shared on Telegram, and spikes occasionally occur simply due to increased activity and discussion between users on antisemitic channels. Furthermore, posts are often reshared multiple times in a short period, or single users post multiple antisemitic posts at once.
67% (106) of the posts containing both antisemitic hate speech and HII concerned globalist conspiracies. There is a strong integration of antisemitic hate speech with globalist conspiracy theories. Antisemitic hate speech therefore followed a similar trajectory over the period as HII – a decrease through December and increase over February. These conspiracies typically frame all Jewish people as involved in a secret global order which aims to subjugate humankind, integrating ideas surrounding financial control, technology and surveillance: “**** control the mainstream media, social-media, banking systems and are in key positions of political power in all White societies. It is very obvious that they are not ‘victims’ - they are a hostile aggressive tribe with an innate biological drive to dominate and subjugate” (08/02/2024).
"There is a strong integration of antisemitic hate speech with globalist conspiracy theories"
10 posts blame Jewish people for the Covid-19 pandemic, framing them as intentionally manufacturing (what users refer to as) the ‘scamdemic’, and creating vaccinations for the purpose of population control. Public institutions such as the World Health Organisation and the World Economic Forum are implicated within this conspiracy of Jewish control, and New Zealand leaders framed as ‘globalist puppets’: “If you are too short-sighted to see that the Jewish takeover worldwide is happening as we speak, and only focus on NZ, UN, WEF-WHO-UN-EU-CCP, etc. but never ever mention Israel and Jewish Supremacy, than all [public figure’s] ‘talks’ are just ‘talks’” (03/03/2024)
4. Anti-Māori Hate Speech and HII
Of the total 2183 posts that were identified as containing hate speech, 478 contained anti-Māori hate, targeting Māori as a social group. Of these, 35 also contained HII, and all 35 contained Māori-specific conspiracy theories.
HII about Māori spiked in mid and late February 2024 on separate occasions. Analysis of the posting data reveals that the majority of posts discuss matters relating to the presence of tikanga Māori in governance and education; “Good on her for making a stand against the **** compulsory Māori indoctrination, it has no place in today's world” (22/02/2024).
A common theme within HII/Māori hate speech posts is the assertion that Māori are much better off in Aotearoa society than other ethnicities, and any disadvantages that Māori face are performative, employed and manufactured by a ‘Māori Elite’ to receive more ‘perks’ from the government. 15 posts also contain HII related to social anxiety and cultural warfare, with one referring to Aotearoa as a “Māori Ethnostate”(27/02/2024).
Another overarching theme holds that Māori are selfish, backwards, and/or violent by nature. Some claim that a so-called “iwi elite” acts largely for itself. Posts promote the idea that British colonisation saved the Māori from themselves. Consequently, some posters argue that Māori wanted to be colonised, and “asked Britain to intervene to become their guardian and protector” (01/02/2024).
For a comprehensive look into the relationship between anti-Māori sentiments and HII, see HEIA’s report on ‘Māori and Harmful Inaccurate Information in Aotearoa’.
5. Anti-immigrant Hate Speech and HII
Anti-immigrant hate speech targets migrants or refugees. Of the total 2183 posts that we identified as containing hate speech, 239 posts contained anti-immigrant hate speech.
Of these, 15 also contained HII, mostly posted throughout late February/early March (see fig. 11). March 6th onwards sees a concentration of anti-immigrant HII which reference a post by a prominent businessman on X: “It is obvious to anyone who is not a fool that this administration is deliberately importing vast numbers of illegals” (X, 2024).
Figure 11. Number of anti-immigrant hate speech posts that also contained HII over time
10 displayed globalist conspiracy theories which suggest a globalist elite is pushing refugees and migrants to ‘invade the nation’: “I don't know how but we need to do more to stop these so-called refugees from coming to our countries. The globalists are moving quicker than most realise, if we don't move we will be lost” (17/02/2024). Despite being posted within Aotearoa, many posts in this category reference migration concerns in other nations - namely the United States, Canada and Europe.
Despite being posted within Aotearoa, many posts in this category reference migration concerns in other nations
This intersection of HII with anti-immigrant hate speech combines nativism – a worldview that sees outsiders as a threat to established communities – with conspiracies that suggest there is a wider scheme being implemented by elites to corrupt and degrade the ‘nation’. Posts claim that this conspiracy between global and national elites will mean that “kiwis will lose control of our country in both a demographic AND ownership sense - a win-win for globalists!” (15/02/2024).
6. Islamophobia and HII
Islamophobic hate speech targets Muslims as a social group. Of the total 2183 posts that we identified as containing hate speech, 72 contained Islamophobic hate speech. 6 of those posts also contained HII sentiments.
4 of the posts contained themes of society anxiety - cultural warfare in particular - and referenced an ‘Islamic takeover’. This form of rhetoric is a hallmark of white supremacist imagination, and conspiracy theories surrounding a transnational ‘white genocide’. One post demonstrates how HII can be employed to cast doubt on real events, and in this case, undermined the atrocity committed against the Muslim community in 2019 by framing the Christchurch massacre as a fabrication: “This is exactly what happened in New Zealand. They imported an Israeli trained Australian terrorist […] and used this event to disarm the population. Police also secretly, under the cover of darkness, turned up on farms and stole their guns, which they need for their jobs” (29/02/2024).
In other cases, Islamophobic hate speech is combined with a range of other prejudices, reflecting a generalised conspiratorial worldview which is common within our dataset. In cases like these, the presence of alternative worldviews is framed as inherently threatening and aggressive: “We will not be silenced! There's a great battle raging in NZ right now! There are so many competing voices” (26/02/2024) - this post goes on to highlight the agendas of untrustworthy politicians and various communities in Aotearoa.
7. Anti-Asian/Southeast Asian Hate Speech and HII
Of the total 2183 posts that we identified as containing hate speech, 143 contained anti-Asian/Southeast Asian hate speech. 11 of these also contained HII.
10 of the posts contained both HII and anti-Asian/Southeast Asian Hate Speech relate to globalist conspiracies, where a common theme found within this overlap is elite control, where China is framed as attempting “to destroy American and Western freedom” (28/02/2024). Sometimes, conspiracy theories surrounding technology and surveillance are combined: “Chinese paying with their hands is death to your freedom and privacy and hello to 5G electromagnetic microwaves radiation cancers 24/7” (07/03/2024)
4 posts combine anti-Asian/Southeast Asian hate speech with HII related to social anxiety and cultural warfare, disseminating ideas surrounding a “*** invasion… operating under the umbrella of the migration”(16/02/2024). Consequently, this sentiment overlaps with anti-immigration rhetoric: “And most kiwis will be bred out by 2050” (29/12/2024).
8. Anti-Black Hate Speech and HII
Anti-black hate speech refers to rhetoric which targets people of colour as a social group, including rhetoric which refers to the white ‘race’ as superior. Of the total 2183 posts that we identified as containing hate speech, we identified 82 posts which contained anti-black sentiments.
7 of these also contained HII, which framed non-white people as manufacturing their experiences of discrimination: “[…] false claims of “racism” from the minorities” (17/02/23). In other cases, racist tropes appear within conspiracy theories which primarily target other communities with posters weighing up which out-group is ‘worse’.
9. Ableist Hate Speech and HII
Ableist hate speech refers to rhetoric which targets disabled or neurodivergent people as a social group. Of the total 2183 posts that we identified as containing hate speech, we identified 92 posts which contained ableist hate speech.
In our dataset, there was only 1 ableist post which also contained HII, and implied that school teachers purposefully give autistic children special treatment by turning a blind eye on any bad behaviour.
Key findings and implications for Aotearoa
Our findings are derived from the analysis of 2183 New Zealand-based hate speech and 3436 New Zealand-based HII posts posted on Facebook, Telegram, Instagram and Facebook over a period of 4 months. These posts were verified as the most harmful examples of HII and hate speech derived from a large collection of 1,546,742 posts collected longitudinally.
We have identified four key findings:
ANTISEMITIC HATE SPEECH OVERLAP WITH HII
Our study found that 27.5% of antisemitic hate speech also contained HII – particularly in conspiracies which centred on elite control. We contend this is because antisemitism has historically included themes of disinformation, and the content of the posts we analysed which implicate contemporary public figures and social issues are the modern manifestation of a longstanding narrative.
SOCIAL GROUPS AS SCAPEGOATS IN HII
Specific forms of hate speech see overlap with specific themes of HII. Antisemitic hate speech strongly correlates with globalist conspiracies, whereas anti-LGBT+ sentiments strongly correlate with social anxiety. In this way, social groups are used as scapegoats for the concerns and anxieties raised within HII posts. More specific research is required to confidently determine if offline events (within Aotearoa or transnationally) which inflame specific themes of HII simultaneously incite spikes in hate towards the correlating community, and vice versa.
HII CONGEALING MULTIPLE ENEMIES (AND ALLIANCES)
HII can tie together different forms of hate speech directed towards different social groups by providing a more ‘coherent’ story and a common enemy ‘outgroup’ made up of multiple social groups. As one example, globalist conspiracy HII brings together antisemitism and anti-LGBT+ sentiments as a combined enemy in one slur. This is particularly clear when HII conspiracies are applied to offline events which implicate multiple social groups. Consequently, hateful HII which does this makes people of those intersecting identities especially vulnerable to isolation, maltreatment, and harm.
It is important to continue pinpointing these unique products which appear with the convergence of HII and hate speech. These combinations of ideas - if taken up by enough people - have the capacity to form into new extreme social movements (Gartenstein-Ross et al., 2023). This was observed in the popularisation of Incel ideas, which have since motivated multiple attacks worldwide, including in Auckland in 2022.
These combinations of ideas - if taken up by enough people - have the capacity to form into new extreme social movements
OVERARCHING THEME: POPULATION AND CULTURAL REPLACEMENT
Population and/or culture replacement/extinction is an overarching theme throughout all hate speech/HII combinations, and regardless of the specific social group targeted, conspiracy theories often demonstrate anxiety related to a changing world. These anxieties are harmful, however, because they also infer that 1) changes are being purposefully implemented in a grander, global, behind-the-scenes scheme; and 2) blame a particular social group(s) for orchestrating these changes.
Population and/or culture replacement/extinction is an overarching theme throughout all hate speech/HII combinations
More specifically, the overlap of HII within the hate speech categories of misogyny, anti-LGBT, anti-Māori and antisemitism often related to concerns around cultural change, replacement/extinction, or corruption. The overlap of HII within the hate speech categories of anti-immigrant, islamophobia, anti-Asian/Southeast Asian, and white supremacy/anti-black centres around migration.
Extremist ideologies are formed on the assertion that the in-group (or its cultural practices) are at threat from some form of intent on the part of the powerful, highlighting the danger of this narrative. In fact, this has been the central idea surrounding several white nationalist terrorist attacks. Tracking the prevalence of hateful posts related to population and/or culture replacement is therefore of crucial importance, as this will identify the social groups used as scapegoats who may become the future victims of ideologically-inspired violence.
Concluding remarks
This report aimed to empirically analyse the relationship between online hate speech and Harmful Inaccurate Information (HII) in Aotearoa utilising online data collected over a period of 4 months. Both hate speech and HII can undermine social cohesion, threaten democracy, and harm target communities. Beyond this, technological advancements which alter how information is produced and consumed has facilitated the globalisation of hate and HII, where events occurring overseas act to fuel their dissemination within Aotearoa. In line with this, the spread and uptake of HII can generate unexpected alliances over a common issue, and string together different forms of hate.
An exploration of online posts which contained both HII and hate speech revealed that, 1) antisemitic hate speech strongly overlaps with HII; 2) social groups are used as scapegoats within HII, 3) HII can converge different strands of hate speech, tying together common enemies and alliances; and 4) an overarching theme throughout all hate speech/HII combinations is population/cultural replacement.
Future research should further track how HII and hate speech interact overtime, and how these interactions respond to and manifest in offline events.
References
AEC. (2023). Disinformation tactics. Australian Electoral Commission. https://www.aec.gov.au/media/disinformation-tactics.htm
Arce-García, S., & Menéndez-Menéndez, M.-I. (2023). Inflaming public debate: a methodology to determine origin and characteristics of hate speech about sexual and gender diversity on Twitter [Article]. El Profesional de la Información, 32(1), 1-18. https://doi.org/10.3145/epi.2023.ene.06
Bundtzen, S. (2023). Misogynistic Pathways to Radicalisation: Recommended Measures for Platforms to Assess and Mitigate Online Gender-Based Violence. https://www.christchurchcall.com/assets/Summit-23/Leaders-Summit-2023-ISD-paper-Misogynistic-pathways-to-radicalisation.pdf
Castaño-Pulgarín, S. A., Suárez-Betancur, N., Vega, L. M. T., & López, H. M. H. (2021). Internet, social media and online hate speech. Systematic review. Aggression and Violent Behavior, 58, 101608. https://doi.org/https://doi.org/10.1016/j.avb.2021.101608
Cinelli, M., Andraž, P., Mozetič, I., Quattrociocchi, W., Novak Petra, K., & Zollo, F. (2021). Dynamics of online hate and misinformation. Scientific Reports (Nature Publisher Group), 11(1). https://doi.org/https://doi.org/10.1038/s41598-021-01487-w
Cosentino, G. (2020). From Pizzagate to the Great Replacement: The Globalization of Conspiracy Theories. In G. Cosentino (Ed.), Social Media and the Post-Truth World Order: The Global Dynamics of Disinformation (pp. 59-86). Springer International Publishing. https://doi.org/10.1007/978-3-030-43005-4_3
Daniel, J. (2018). the algorithmic rise of the “alt-right”. Contexts, 17(1), 60-65. https://www-jstor-org.ezproxy.auckland.ac.nz/stable/26394263
EFSAS. (2021). The role of fake news in fueling hate speech and extremism online. https://www.efsas.org/publications/study-papers/the-role-of-fake-news-in-fueling-hate-speech-and-extremism-online/
Faulkner, N., & Bliuc, A.-M. (2016). ‘It’s okay to be racist’: moral disengagement in online discussions of racist incidents in Australia. Ethnic and Racial Studies, 39(14), 2545-2563. https://doi.org/10.1080/01419870.2016.1171370
Finkelstein, J., Paresky, P., Goldenberg, A., Zannettou, S., Jussim, L., Baumgartner, J., Riggleman, D., Farmer, J., Goldenberg, P., Donohue, J., & Modi, M. H. (2020). Antisemitic Disinformation: A Study of the Online Dissemination of Anti-Jewish Conspiracy Theories (Contagion And Ideology, Issue. https://networkcontagion.us/reports/antisemitic-disinformation-a-study-of-the-online-dissemination-of-anti-jewish-conspiracy-theories/
Flora, T., & Jemma, C. (2020, October 13). Hate crime reports to police hit record high. Press Association. https://ezproxy.auckland.ac.nz/login?url=https://www.proquest.com/wire-feeds/hate-crime-reports-police-hit-record-high/docview/2538016863/se-2?accountid=8424
Gartenstein-Ross, D., Zammit, A., Chace-Donahue, E., & Urban, M. (2023). Composite Violent Extremism: Conceptualizing Attackers Who Increasingly Challenge Traditional Categories of Terrorism. Studies in Conflict & Terrorism, 1-27. https://doi.org/10.1080/1057610X.2023.2194133
Gibbons, A., & Carson, A. (2022). What is misinformation and disinformation? Understanding multi-stakeholders’ perspectives in the Asia Pacific. Australian Journal of Political Science, 57(3), 231-247. https://doi.org/10.1080/10361146.2022.2122776
Ging, D. (2017). Alphas, Betas, and Incels: Theorizing the Masculinities of the Manosphere. Men and Masculinities,22(4), 638-657. https://doi.org/10.1177/1097184X17706401
Hameleers, M., van der Meer, T., & Vliegenthart, R. (2022). Civilized truths, hateful lies? Incivility and hate speech in false information – evidence from fact-checked statements in the US. Information, Communication & Society,25(11), 1596-1613. https://doi.org/10.1080/1369118X.2021.1874038
HEIA. (2023). Disinformation Trends in New Zealand. U. o. Auckland. https://www.heiaglobal.com/post/disinformation-trends-in-new-zealand
HEIA. (2024). Harmful Inaccurate Information: Defining disinformation, misinformation - what it is (and what it is not). U. o. Auckland. https://www.heiaglobal.com/post/harmful-inaccurate-information-defining
Herz, M., & Molnar, P. (2012). The Content and Context of Hate Speech : Rethinking Regulation and Responses. Cambridge University Press. http://ebookcentral.proquest.com/lib/auckland/detail.action?docID=880678
Johnson, N. F., Leahy, R., Restrepo, N. J., Velasquez, N., Zheng, M., Manrique, P., & Devkota, P. (2019). Hidden resilience and adaptive dynamics of the global online hate ecology [Report]. Nature, 573, 261+. https://link.gale.com/apps/doc/A602150517/AONE?u=learn&sid=bookmark-AONE&xid=ec90b15e
Langer, A. (2022). Deep State, Child Sacrifices, and the “Plandemic”: The Historical Background of Antisemitic Tropes within the QAnon Movement. In M. Hübscher & S. Von Mering (Eds.), Antisemitism on Social Media. Routledge. https://doi-org.ezproxy.auckland.ac.nz/10.4324/9781003200499
Leitch, S., & Pickering, P. (2022). Rethinking Social Media and Extremism. ANU Press. http://ebookcentral.proquest.com/lib/auckland/detail.action?docID=7023487
Lupu, Y., Sear, R., Velásquez, N., Leahy, R., Nicholas Johnson, R., Goldberg, B., & Johnson, N. F. (2023). Offline events and online hate. PLoS One, 18(1). https://doi.org/https://doi.org/10.1371/journal.pone.0278511
Manne, K. (2017). Down Girl: The Logic of Misogyny. Oxford University Press. https://doi.org/10.1093/oso/9780190604981.001.0001
Marwick, A., & Lewis, R. (2017). Media Manipulation and Disinformation Online. https://datasociety.net/library/media-manipulation-and-disinfo-online/
Milhazes-Cunha, J., & Oliveira, L. (2023). Doctors for the Truth: Echo Chambers of Disinformation, Hate Speech, and Authority Bias on Social Media. Societies, 13(10), 226. https://www.mdpi.com/2075-4698/13/10/226
Mirchandani, M. (2018). Digital Hatred, Real Violence: Majoritarian Radicalisation and Social Media in India. https://www.orfonline.org/wp-content/uploads/2018/08/ORF_OccasionalPaper_167_DigitalHatred.pdf
Robie, D. (2022). FRONTLINE 1: New Zealand’s 23-day Parliament siege: QAnon and how social media disinformation manufactured an ‘alternate reality’. Pacific Journalism Review : Te Koakoa, 28(1 & 2), 105-113. https://doi.org/10.24135/pjr.v28i1and2.1265
Searles, M. (2024, 18 February ). Trans-women’s milk as good as breast milk, says NHS trust. Telegraph. https://www.telegraph.co.uk/news/2024/02/18/trans-womens-milk-as-good-as-breast-milk-says-nhs-trust/#:~:text=In%20a%20letter%20to%20campaigners,the%20birth%20of%20a%20baby%E2%80%9D.
Shand, A. O. M. (2021, November 1). Hate Crimes in New Zealand: Why Do We Know So Little? Equal Justice Project. https://www.equaljusticeproject.co.nz/articles/hate-crimes-in-new-zealand-why-do-we-know-so-little2021
Udupa, S., Gagliardone, I., Deem, A., & Csuka, L. (2020). Hate Speech, Information Disorder, and Conflict: Research Review.
Vasist, P. N., Chatterjee, D., & Krishnan, S. (2023). The Polarizing Impact of Political Disinformation and Hate Speech: A Cross-country Configural Narrative. Information Systems Frontiers. https://doi.org/10.1007/s10796-023-10390-w
Wang, Y., Bye, J., Bales, K., Gurdasani, D., Mehta, A., Abba-Aji, M., Stuckler, D., & Martin, M. (2022). Understanding and neutralising covid-19 misinformation and disinformation. BMJ : British Medical Journal (Online), 379. https://doi.org/https://doi.org/10.1136/bmj-2022-070331
Williams, M. L., Burnap, P., Javed, A., Liu, H., & Ozalp, S. (2019). Hate in the Machine: Anti-Black and Anti-Muslim Social Media Posts as Predictors of Offline Racially and Religiously Aggravated Crime. The British Journal of Criminology, 60(1), 93-117. https://doi.org/10.1093/bjc/azz049
Wilson, C., Renner, E., Smylie, J., & Dziwulski, M. (2024, March 14). Christchurch attacks 5 years on: terrorist’s online history gives clues to preventing future atrocities. The Conversation. https://theconversation.com/christchurch-attacks-5-years-on-terrorists-online-history-gives-clues-to-preventing-future-atrocities-225273
X. (2024, March 6). It is obvious to anyone who is not a fool that this administration is deliberately importing vast numbers of illegals. No grand conspiracy theory is needed to explain this, just simple incentives: they are far more likely to vote Democrat than Republican. X (Twitter).
Xia, L. (2024, February 21). Rise in hate incidents reported to police.Radio New Zealand.https://www.rnz.co.nz/news/national/509717/rise-in-hate-incidents-reported-to-police
For enquires regarding this report please contact Dr. Chris Wilson at chris@heiaglobal.com
Comments