X released today the company’s first transparency report since Elon Musk bought the company, formerly Twitter, in 2022.
Prior to Musk’s takeover, Twitter would publish transparency reports every six months. They largely cover the same ground as the new X report, giving specific download numbers, government requests for information and content removals, and data on what content has been reported and, in some cases, removed for violating the rules. The most recent transparency report available from Twitter covered the second half of 2021. and it was 50 pages long. (X is shorter than 15 pages, but requests from governments are also listed elsewhere on the company’s website and are consistently updated to stay compliant with various government orders.)
Comparing the 2021 report with the current transparency report X is a bit difficult because the way the company measures different things has changed. For example in 2021 11.6 million accounts were reported. Of those 11.6 million, 4.3 million were “triggered” and 1.3 million were suspended. According to the new X report, there have been over 224 million reports of both accounts and individual pieces of content, but the result has been 5.2 million accounts that have been suspended.
While some numbers remain seemingly consistent across reports—reports of abuse and harassment are somewhat predictably high—in other areas there is a huge difference. For example, in the 2021 report accounts reported for hateful content accounted for nearly half of all reports, and 1 million of the 4.3 million accounts took action. (The reports were previously interactive on the website; the current PDF no longer allows users to scroll through the data for more detailed breakdowns.) In the new report, X says the company took action against only 2,361 accounts for posting hateful content.
But that could be because X’s policies have changed since it was Twitter, which Theodora Scheadas, a former member of Twitter’s public policy team who helped assemble the moderation research consortium, says could to change the way numbers look in a transparent report space. For example, last year the company changed its policies on hate speech, which previously covered misgendering and name-calling, and lifted its policy on misinformation about Covid-19 in November 2022.
“Because some policies were changed, some content is no longer in violation. So if you’re looking at changes in the quality of experience, that can be difficult to capture in a transparency report,” she says.
