I’ve been running my own very unscientific experiment on X over the past couple months. Almost every time I see a post from Elon Musk, I click the three dots in the upper right corner and select “not interested in this post.” The question I’m trying to answer: will X’s algorithm actually stop showing me the uninformed bile spewed by its owner?
So far, after months of clicking, Musk’s posts inevitably return to the top of my feed…
Perhaps this shouldn’t be surprising. As The Verge reported back in February of 2023, earlier that month, James Musk, Elon’s cousin and an X employee (because of course he is), sent a “high urgency” message to the company’s engineers. Elon’s post about the Super Bowl had received fewer impressions than one sent by then President Joe Biden. Within hours, X’s algorithm had been altered to boost Musk’s tweets by a factor of 1,000.
Did Elon Musk buy Twitter and transform it into X just to force its users to witness his childish, ketamine-fueled stream of consciousness, which manifests across nearly 70 posts per day? If so, he’ll likely never admit it.
Musk’s publicly stated reasons for buying Twitter were to make the platform a bastion of free speech and to liberate it from the bots that he claimed had overrun it. Sounds like social media utopia! But according to a genuinely scientific analysis published earlier this month, the reality has been far more dystopic.
That analysis, conducted by researchers at UC-Berkeley and the University of Southern California, showed that hate speech on X increased 50 percent since Musk’s purchase. In particular, transphobic slurs leapt 260 percent, homophobic tweets rose 30 percent, and racist posts jumped 42 percent. At the same time, there was no significant reduction in activity of “inauthentic accounts.” In fact, bot activity related to cryptocurrency actually increased.
The study looked at English-language posts on the platform between January 2022 to June 2023. Obviously, a lot has happened since then. Has X become even more of a cesspool in the last year and a half or have its waters cleared up?
Data that X released in its latest global transparency report do not paint a clear picture. The platform says it is suspending more accounts than ever for violations of its rules – 5.3 million in the first half of 2024 versus 1.6 million in the first half of 2022 before Musk’s takeover. The increase in suspensions could be due to greater enforcement or perhaps a surge in malicious accounts – it’s impossible to tell from the data.
What is clear from the data is that X is suspending vastly fewer accounts for “hateful conduct” – 111,056 in 2022 versus only 2,361 in 2024 – which hints that the noxious trend identified by the Berkeley and USC team likely has continued.
Timothy Graham, an Associate Professor in Digital Media at the Queensland University of Technology, and Mark Andrejevic, a Professor in the School of Media, Film, and Journalism at Monash University turned up additional evidence that X has remained dodgy. The duo’s working paper, quietly released in November of last year, showed that almost exactly at the time Elon Musk endorsed Donald Trump for president on July 13, 2024, there was a “structural engagement shift” on the platform in a couple key areas.
“Musk’s account not only started with a higher baseline compared to the other accounts in the analysis but also received a significant additional boost post-change, indicating a potential algorithmic adjustment that preferentially enhanced visibility and interaction for Musk’s posts.” they reported.
At the same time, Republican-leaning accounts exhibited a significant increase in views relative to Democrat-leaning accounts.
“This finding suggests a possible recommendation bias favouring Republican content in terms of visibility,” they reasoned.
A Washington Post analysis found something similar. Between July and December of 2023, eight tweets from Democrats in Congress had more than 20 million views, compared to five from Republicans. Between July and October of 2024, Democrats had six of these viral tweets, while Republicans had 29.
In short, it seems that free speech “champion” Elon Musk has been tipping X’s scales in favor of his posts and of those sharing his political ideology. Is speech on X really “free” if its owner secretly boosts speech he likes and muffles speech he doesn’t?
From DOGE to X, Musk claims to be a champion of transparency, but his actions suggest otherwise. Daniel Hickey, first author on the aforementioned hate speech paper, told Splinter that X cut off academics’ free access to its API in mid-2023. This is the system that researchers used to collect data from the platform.
“This means that we would have to pay a hefty sum to keep monitoring trends in hate speech and bots, and this wasn’t feasible for us to do,” he said in an email. “Not only does this policy change of reduced access to data mean the platform is less transparent and it is more difficult to hold them accountable, but it is detrimental to social science research as a whole, as Twitter/X were previously a rich and widely accessible data source for researchers to answer all kinds of questions about human behavior on social media platforms.”
He added that Reddit is currently beta testing a program that will give researchers easy access to data from their platform, and X-competitor Bluesky already grants free access.
Since Musk mutated Twitter into X, the platform is now more hateful, less transparent, just as polluted with bots, and even more polluted with Musk. The other day, the world’s richest man unwantedly crept back to the top of my feed again. This time, I decided to end the experiment. Click the three dots… Scroll down… “Block @elonmusk.”
GET SPLINTER RIGHT IN YOUR INBOX
The Truth Hurts
