No one disagrees with Elon Musk about the success of Twitter moderation
Elon Musk hailed the success of his moderate policies in early December. However, many signals contradict this claim.
The hate is back on Twitter. Or rather, he expresses himself there more comfortably than before. This is the result of the work carried out in the United States, which was repeated by the New York Times on December 2, 2022. The result of the rebellion of the new owner of Twitter, Elon Musk, to call him” completely wrong ».
After the American businessman took over the social network, observations by several organizations highlight a sharp increase in anti-Semitic messages (a jump of more than 60% two weeks after the acquisition of Twitter). As for homophobic content, their daily appearance has increased by more than 58%.
Even research in the field of counter-terrorism shows that the community site is very clearly relaxed. In the first twelve days after Elon Musk took office, 450 accounts controlled by the Islamic State organization were discovered. This represents a 69% increase over the previous twelve days.
These various analyzes were provided by the Center to Fight Digital Hate, the Anti-Defamation League and the Institute for Strategic Dialogue. ” Elon Musk has sent a Bat Signal to all racists, misogynists and homophobes that Twitter is available. They reacted accordingly “The head of the Center for Combating Digital Hate, Imran Ahmed commented.
Elon Musk is happy with Twitter’s moderation policy
Elon Musk’s policy also has a positive effect on the increase of illegal speech. He saw fit to pardon a large number of banned accounts based on an online poll — instead of using the moderation board he promised to put in place. More than 12,000 profiles, including neo-Nazis and anti-Semites, have been reactivated.
An article was published in the New York Times that day, Elon Musk announced his results, suggesting its decisive effect on moderation. In the timeline from October 17 to November 30, 2022, hateful content exposure would drop to lower levels than previously seen after a spectacular peak.
” The number of hate speech impressions (the number of times a tweet is seen) continues to decrease, despite a significant increase in the number of users! […] Freedom of speech does not mean freedom of speech. Negativity should and will reach less than positivity “, he said. In his opinion, hateful content is serious less than 0.1% of total impressions. Surprisingly, while the amount of hateful content is increasing, impressions of this speech are decreasing. Elon Musk does not give more details.
The limits of Elon Musk’s strategy became clear in a sequence around rapper Kanye West. Although he described himself as a “free speech absolutist” a few months ago, he has made several anti-Semitic messages that have prompted pressure from Elon Musk. It is not the first time that the singer has been kicked out of the platform.
Twitter is less quick to respond to reports
Disadvantages of moderation on Twitter is also highlighted by the European Commission. If Brussels is seeing a general decline in mobilization among platforms that have signed the EU Code of Conduct (YouTube, Facebook, Instagram, etc.), it is Twitter that “ marking the biggest drop in a year “.
As part of this code of good conduct, platforms are required to process reports of hate messages within 24 hours. In 2020, partner sites managed to process 90.4% of notifications on time. This statistic dropped to 81% in 2021 and 64.4% in 2022. Only TikTok remains on a positive curve.
The collapse of moderation on Twitter is partly due to the social network’s dramatically reduced ability to have enough staff to manage reports. In mid-November, we learned that contracts with thousands of providers were terminated, and the platform was deprived of several thousand moderators.
On November 30, Twitter gave a very different speech: its team promises to continue like this ” its diligent work to protect the platform from hateful behavior, abuse, and any violation of Twitter’s rules. The team remains strong and well-resourced “. However, it was also noted that the service will increasingly rely on automated detection to handle moderation. Obviously, the results are not yet in place.