Memes of a little boy diligently copying and copying the homework of the little boy next to him, X is full of them at the moment. They refer to Mark Zuckerberg, the boss of Meta (Facebook, Instagram and Threads), who announced on Tuesday that Meta will stop using external fact checkers and moderators. He is following in the footsteps of Elon Musk, who started this earlier on social media platform X.
What exactly does X’s method entail? And what does it mean that Meta is now taking over? Eight questions about how X functions and what path Meta is taking now.
1
So what did Musk do on X?
When Musk took over Twitter more than two years ago and turned it into X, he made some radical changes. As a result, the company not only had to become much more efficient and innovative, but above all much less political. “Censoring,” as Musk calls it, had to end. A term that Donald Trump also often uses and that Zuckerberg also used a number of times on Tuesday.
People whose X accounts had been blocked, including Donald Trump himself, were immediately given back access under Musk. He also fired the teams tasked with assessing whether content crossed legal and moral boundaries.
The experience at X has changed considerably since then. Less controversial content is banned, such as potentially discriminatory comments and propaganda.
2
Is that bad?
You may have different opinions about that. Some people believe that X has become freer and has therefore improved. Others say that it has become a kind of confusing open sewer and therefore no longer a place for funny, fast and topical debate.
Ultimately, it is the user numbers that count and we know little about that for sure. X is a company and the number of (active) users is competitively sensitive information.
According to self-reports to the European Commission, X has more than 100 million users in the EU. Taken together, the various Meta platforms have approximately 259 million users in the EU.
3
Is nothing checked at X anymore?
Yes. X has – just like other major online platforms – systems that automatically filter for prohibited expressions, such as the distribution of child abuse material or incitement to terrorism. These systems (algorithms) can automatically make posts untraceable and block accounts. This happens with millions of posts per day.
In exceptional cases, in case of doubt or complaints, another person will look at it.
4
And what about more difficult categories such as misinformation (incorrect messages), disinformation (intentionally incorrect messages) and, for example, racism?
Previously, people were hired for this, who could intervene, for example by sending a message viral was utter and dangerous nonsense. Difficult work, because there is a large gray category. For this reason, both Twitter and Meta repeatedly came under fire during the Covid pandemic and were accused of censorship.
Immediately after acquiring Twitter, Musk fired about 4,400 employees who did this work. It is not entirely clear how many there were before that time. However, the dismissal meant an explicit change in policy.
Now there are no longer people who constantly actively monitor the content and keep an eye on whether misinformation is being spread.
That system has been replaced by what X calls ‘Community Notes’. This means that a select group of volunteers can attach a note with context to posts from other users, to make it clear that the content is questionable. Such a note must have a source reference.
Whether a note is actually added is decided by an algorithm. This would be programmed in such a way that it always ensures that different perspectives have been weighed.
5
How many moderators are left at the major social media companies?
The main source of information for these types of numbers are the transparency reports that major online platforms must publish according to the European Digital Services Act (DSA). This has made a little more information available so that the companies can be compared with each other and about the operation of their algorithms.
That shows that X currently has 1,275 people working on content moderation worldwide. Until the end of 2023, one of them spoke Dutch, but in May 2024 this was also deleted.
Meta currently has about 40,000 people working on it ‘trust and security’. About 15,000 of them are specifically involved in assessing content, 54 of which are for the Dutch language area.
6
Is Community Notes an effective system?
Moderate. Many examples have been documented over the past two years showing that the system is too slow to stop the spread of disinformation and misinformation, for example about the war in Gaza. A large proportion of misleading posts do not receive a ‘note’. The messages that do receive this have already been shared so many times and therefore spread so far that the note with context no longer makes much sense. Random samples show that the quality of the notes and source references is good.
7
How has Meta dealt with content checks in recent years?
It has proven to be a long-term struggle. The underlying question for all online platforms is: to what extent are they merely neutral conduits, and when are they also responsible for the content they help distribute? Questions that publishers of ‘traditional’ media are familiar with.
After the 2016 elections (in which Trump came to power for the first time), Facebook received a lot of criticism due to the amount of misinformation and disinformation that was spread on the platform during the election campaign. In response, the company set up an extensive system of moderation. It also started using third-party fact-checkers, including international news agencies such as AP and AFP.
The work of moderators quickly came under fire. They turned out to be nasty underpaid jobs. While these are complex decisions, sometimes with a lot of impact for those involved. Right-wing and conservative opinion makers in particular have complained in recent years that their opinions were actually censored under the guise of safety and decency.
Meta was constantly making adjustments. For example, an oversight body (Oversight Board) was established of leading external experts, who subsequently ruled on controversial decisions, such as blocking accounts of public figures. And the algorithms are constantly being adjusted. Then more news and current discussions, then more children’s photos and cat videos.
8
Do Zuckerberg’s announcements also apply to users in Europe?
Zuckerberg emphasized in his message that the changes are initially only for the US. However, on Wednesday it emerged that the company also wants a different approach in the EU. It has sent intentions in this regard to the European Commission (EC), which has yet to respond, NOS reports.
The legislation in the US is very different than in the EU. The EU has had the Digital Services Act since 2022. Under this law, the large online platforms in particular (with more than 45 million users per month in the EU) have been given a lot more obligations and responsibilities. Not so much for what users publish, but for the way in which companies deal with it. Because it makes a big difference whether you let someone honk something in a corner, or whether you give that person a megaphone if it turns out that people like to click on it.
The DSA demands transparency about the operation of the algorithms. And responsible use of it. So, for example, not to commercially exploit children and bombard them with advertisements aimed at them. The EC has started an investigation against several platforms due to suspected violations of the DSA. Below that X and Meta.
The EU hopes that the legislation will be copied in other parts of the world. This makes it easier to enforce globally operating companies.
Zuckerberg also used his video message to make a kind of declaration of war against other jurisdictions, including the EU, which he characterizes as enemies of freedom of expression.
“Europe has an ever-growing number of laws that institutionalize censorship and make it difficult to build anything innovative there.”
Meta wants to fight against this together with the new American government. So it does not look like the US will copy European legislation in the short term.

