In the days after Elon Musk took over Twitter in October 2022, the social media platform saw a “peak of hateful actions”, which the safety leader later put down to a “focused short-term trolling campaign.” A new study shows that when it comes to antisemitism, there is nothing.
However, antisemitic tweets have doubled in the months since Musk took office, according to research conducted by myself and colleagues at the technology company CASM Technology and the Institute for Strategic Dialogue think tank. Between June and October 26, 2022, the day before the acquisition of Twitter by Musk, there was a weekly average of 6,204 tweets that were considered “possible antisemitic” – that is, at least one reasonable interpretation of the tweet was in the International Holocaust Remembrance Alliance. The definition of the term is “a certain perception of Jews, which can be described as hatred of Jews.”
But from October 27 to February 9, 2023, the average was 12,762 – a 105% increase. In total, a total of 325,739 tweets from 146,516 accounts were labeled as “possible antisemitic” during the study, from June 1, 2022 to February 9, 2023.
Discovering antisemitism with AI
To identify antisemitic tweets, my co-authors and I combined 22 published hate speech identification algorithms into a single mechanism and used more machine learning to see which combinations of decisions led to the correct results. We then went through all the tweets – more than a million in total – that contained one of 119 words, phrases, slurs and epithets related to antisemitism.
No process is perfect. We estimate our model makes the correct decision about 75% of the time. We also missed some antisemitic tweets that did not contain those 119 keywords, as well as those captured before early December when we collected the data.
We then used an algorithm to draw 10 different themes of antisemitism that appeared in the tweets. Some centered around the use of certain antisemitic insulting epithets. Others cite conspiracy theories about hidden Jewish influence and control.
The antisemitic tweets directed at Jewish investor and philanthropist George Soros warrant a category of their own. He was mentioned more than anyone else in our data, more than 19,000 times, with tweets claiming he was a hidden globalist, Jewish or “Nazi” member of the world order.
Another theme was a tweet defending the rapper Ye, the predecessor of Kanye West, who had made some antisemitic comments after his account was reinstated by Musk.
Our research, which has not been peer-reviewed, also found around 4,000 antisemitic tweets focused on the Russian invasion of Ukraine. This variety claims that the conflict was caused by Jews, or that Jews secretly caused the US to support Ukraine. They also contain direct antisemitism aimed at the president of Ukraine, Volodymyr Zelenskyy, who is Jewish.
Musk is rolling back content moderation
Musk’s acquisition of Twitter is behind what I see as a decades-long trend among tech giants to take more responsibility for hate speech, harassment, incitement, disinformation and other harm contained in the information flowing through the platform. During this period, companies such as Facebook and Twitter gradually implemented policies to respond to extremism, hate speech and harassment, or increase “civility,” as Twitter itself explained in 2018, and created teams and tools to implement them.
Musk, who claims to be “absolutely free speech,” pointed the platform in a different direction after taking control. In a short period of time, Twitter’s Trust and Safety Council was disbanded, previously banned accounts were reinstated and more than half of Twitter’s staff were fired or simply left — including many responsible for enforcing the company’s hate speech policy.
As someone who has tracked hate speech on places like Twitter for about 10 years, I believe changes in Twitter’s moderation practices are only partially to blame for the rise of antisemitism on the platform.
The media spectacle surrounding Musk’s takeover, along with his vocal views on free speech, may also encourage them to join or rejoin platforms that have failed from previous attempts to confront hate. Our research provides some backing for this theory. About 3,855 accounts identified as having posted at least one antisemitic tweet joined Twitter in the 10 days after Musk took over. However, this is only a small fraction of the 146,516 accounts that posted at least one antisemitic tweet during the entire study.
Small effect on preventing hate speech
The rise of hate speech on Twitter has been marked by researchers in the weeks after Musk took over, concerned with the billionaire. fired as “very false,” before swearing “max deboosted & demonetized“hateful tweets.
If Twitter is de-amplifying antisemitism, our research shows almost no evidence of that. Before October 27, antisemitic tweets received an average of 6.4 “favorites” and 1.2 retweets. Since then, they have averaged 6 “favorites” and 1 retweet. Although engagement is not a perfect measure of visibility, tweets that are less visible to users will generally get less engagement.
We also tried to measure takedowns of antisemitic tweets. On February 15, 45 days after we first collected data, we tried to collect again all the tweets we identified as antisemitic. Tweets may be unavailable for many reasons, and Twitter enforcement is just one of them. It’s not perfect though, it gives us a tentative glimpse of what can happen in regards to removing antisemitic posts. And on that date, 17,589 antisemitic tweets were taken down – 8.5% of the total.
Rising tide of antisemitism
Our findings come at a time when many fear a threat to the Jewish community. In 2021, the Anti-Defamation League tracked the most antisemitic incidents – including harassment, vandalism and assault – in the US since it started tracking the numbers in 1979. And this is not just a US phenomenon; in England, the Community Security Trust has recorded a similar surge in anti-Jewish activity, while in Germany, anti-Jewish crimes have risen by 29% due to the pandemic.
Studying social media has shown me time and time again how powerfully it helps shape the culture and ideas that underlie user behavior. Ultimately, the proliferation of tweets that hold Jews responsible for all the world’s ills, that propagate control conspiracies and cover-ups, or fire attacks directed at Jews, can only support antisemitism online – and in the real world.
Carl Miller is Research Fellow, King’s College London.
This article is republished from The Conversation under a Creative Commons license. Read the original article.