Study finds durable rise in online religious hate in Australia

A new Tackling Hate Lab study of more than 2 million Australian social media posts found anti-Jewish and anti-Muslim abuse online rose sharply after Oct. 7, 2023, and stayed above earlier levels through the research period. The findings add evidence to Australia’s current policy debate over hate, as a Royal Commission into Antisemitism and Social Cohesion proceeds and federal special envoys work on antisemitism and Islamophobia.

The research group released two reports on the issue, according to an account in The Conversation by researchers Matteo Vergani, Andrea Giovannetti and Kewen Liao. The team examined Australian social media posts, most from X, formerly Twitter, along with hundreds of verified offline incidents involving vandalism, harassment and physical violence against Jewish and Muslim Australians from 2021 to 2026.

How the researchers measured hate

The researchers said they tested several definitions of anti-Jewish and anti-Muslim hate because public debate often turns on where those lines are drawn. Across the reports, they used Google’s Perspective API to flag identity attacks, which they described as language that insults, demeans or dehumanizes people based on identity.

The team then used the large language model Qwen to identify the target of the abuse, including Jews, Israelis, Zionists, Muslims and Palestinians, according to The Conversation. The researchers also used an AI model built to reflect how Australian Jewish and Muslim community members perceive anti-Jewish and anti-Muslim hate online.

The authors said the size of the measured increase differed by method, but the direction of the trend did not. They also cautioned that the anti-Jewish and anti-Muslim datasets were assembled differently, so the reports support comparisons over time within each dataset rather than a direct comparison of which type of hate was more common overall.

Oct. 7 changed the baseline

Identity attacks aimed at Jews rose from an average of 0.3 posts a day in the year before Oct. 7, 2023, to 16.8 posts a day in the following year, the researchers reported. Attacks aimed at Israel increased from 1.3 to 78.9 posts a day over the same period.

Anti-Muslim identity attacks also rose after Oct. 7, increasing from 2.8 posts a day beforehand to 42 posts a day in the following year, according to the reports. The largest anti-Muslim surge came after the Bondi terrorist attack, when posts targeting Muslims averaged 1,323.3 a day in the week after the attack and then settled at about 224.4 a day over the next month.

The researchers said the same broad pattern appeared when they looked at identity attacks, toxic language and different target groups. They said Oct. 7 raised the baseline for online hate, while Bondi produced a sharp anti-Muslim surge and a smaller anti-Jewish spike.

Online and offline incidents became linked

The reports also found a closer connection between online hostility and offline hate incidents after Oct. 7. Before that date, the researchers said online discussion and offline incidents were often less connected.

Afterward, vandalism, harassment and physical attacks targeting Jewish or Muslim people and property were more likely to be followed by hostile online activity, according to the reports. In the anti-Muslim dataset, each verified offline incident was associated with an average of about 12 additional hateful posts.

The researchers said news coverage of real-world incidents could create openings for a minority of online users to spread abuse, conspiracy theories and hostility in comments and social media posts. They argued that governments need longer-term evidence collection, including a permanent Observatory of Hate tracking online and offline abuse against Jewish and Muslim Australians as well as First Nations people, migrants, women, people with disabilities and people of diverse genders and sexualities.

This story draws on original reporting from Phys.org.