“This is how Musk encouraged state disinformation about X”: the NewsGuard report

In recent years, the use of social media as a political propaganda tool has taken on an unprecedented role in the global political landscape. These online channels have transformed into powerful tools of persuasion and influence, allowing politicians to reach large audiences immediately and effectively. The strategic use of social media in politics – above all Facebook, X (formerly Twitter), Instagram and YouTube – has already influenced elections, public opinions and political debates in many parts of the world. Political propaganda on social media has evolved through several phases, from targeted content creation and audience segmentation to the use of algorithms to amplify specific messages. Political leaders have acquired the ability to communicate directly with voters, bypassing traditional media channels and establishing immediate contact with their supporters, often thus eliminating ‘inconvenient’ intermediaries.

NewsGuard’s analysis of X after the policy change

This evolution has raised important ethical questions and raised concerns about disinformation, manipulation of public opinion and user privacy, pushing platforms to self-regulate to limit the spread of fake news as much as possible. Analysts at NewsGuard Technologies, a New York-based global media and information monitoring company, have been focusing on the use of X, the platform formerly known as Twitter, particularly since Elon Musk, who bought the social network in 2022, has decided to change the policy regarding the clear indication to users of accounts affiliated with or controlled by governments.

Profiles of Russian, Chinese and Iranian origin monitored

More specifically, NewsGuard experts conducted a comparison of engagement data (i.e. the number of likes and shares) relating to 12 English-language media accounts associated with governments from Russia, China and Iran , in the 90 days before and after the policy change. The findings highlight how the profiles of these state sources now have the ability to reach a larger and potentially more defenseless audience, as users may not be aware of the main intent of these sources: the dissemination of propaganda.

With the change in policy, engagement increased by 70% for Russian, Chinese and Iranian accounts

According to the NewsGuard report, which is based on data from media monitoring platform Meltwater, in the three months following the removal of labels indicating state affiliation or funding of an account, the engagement generated by Russian state media , Chinese and Iranians on X increased by 70%. In particular, RT – which used to be called Russia Today, but changed its name several years ago – recorded the highest level of engagement after users could no longer identify it as a Kremlin-affiliated media outlet, almost doubling it and “going from 1, 3 million likes and reposts when the label was still available at 2.5 million after its removal,” as stated in the document.

After the policy change, Tass, another Russian agency, also saw a 63% increase in engagement, while Iran’s PressTV had 97% and China’s Global Times 26%.

Although profiles monitored by NewsGuard had published approximately the same number of posts (as “tweets” on X are now called) compared to the previous 90-day period, government disinformation sources all saw increased engagement. “In the three months following X’s decision to remove the labels – the report continues – the 12 accounts (four Russian, four Chinese and four Iranian) monitored obtained a total of 4.98 million likes and reposts out of a total of 63,108 post. In the 90 days prior to the change, accounts had received 2.93 million interactions on 62,551 posts.” NewsGuard said Musk’s platform did not respond to two emails and a message about state-controlled media, a decision never explained by the tycoon’s company. Musk’s only mention of the topic came in a post from April 2023, the same month the new policy was introduced, where he wrote that “all news is propaganda to some extent” and that people should “decide Sunglasses”.

How X worked before Musk’s changes

The decision to remove the reports was taken 16 days after a dispute between While receiving public resources, NewsGuard points out that NPR works independently and also receives money from “programming fees paid by member stations, which are supported in part by the federal government and state and local governments.” Before Musk’s changes to the platform’s policies, X provided users with information about an account’s government affiliation both in the magazine’s profile and on each individual message. Additionally, posts from any account that contained links to state media articles were marked with an orange exclamation mark and the words ‘Stay informed’. “It is now impossible for users to know whether or not an account is affiliated with a government unless they already know or do independent research,” NewsGuard experts point out.

In its study, NewsGuard also notes that none of the 12 state accounts it monitored – which together have 30.9 million followers – “voluntarily discloses government affiliation.” For example, the Russian state news agency TASS describes itself in X like this: “With us, news becomes stories and events become history.” The Russian newspaper RT instead states: “Freedom wins over censorship, truth over storytelling”, and the account carries the wording “Media & News Company”. Although these 12 state accounts on X are known for spreading disinformation, the analysis revealed that, in the three months following Musk’s policy change, “the posts with the highest engagement did not contain blatantly false information. The best-performing posts, however, regularly included memes and propaganda ostensibly aimed at undermining the West and his allies.”

This is how the use of memes promotes propaganda

To support its thesis, NewsGuard refers to the Iranian account of Press TV, which on July 1, 2023 shared a meme that joked about French President Emmanuel Macron’s response to protests against police violence in France and demonstrations for women’s rights in Iran, accusing Macron of having “double standards”. This post, which received over 8,000 likes and reposts as of September 20, 2023, was the third most popular on the account in the three months after the labels were removed. Michael Lynch, a philosophy professor at the University of Connecticut and author of a study published in April 2022 on the function of memes as a vehicle for the spread of disinformation, explained to NewsGuard that government-affiliated profiles use memes for propaganda because they manage to “sweeten” the message. The expert also noted that many of these profiles try to make people laugh, as humor tends to put people at ease and “makes them more open to the propaganda message.”

However, the removal of labels was not the only change that state media profiles benefited from. “Another reason why these accounts may have seen an increase in engagement – ​​the report continues – is that users no longer have to actively search for content from state media outlets. NewsGuard has indeed discovered, thus confirming the conclusions of other researchers, that the X algorithm now regularly promotes this type of content”. From August 2020 to early 2023, X (then still Twitter) had banned the promotion of government-affiliated accounts, stating that it “expanded the reach of state-affiliated media accounts or their tweets through our recommendation systems, presenting them on the home page, among notifications and as search results”. After Musk’s changes, the press documented how state media accounts are now regularly recommended by the algorithm on users’ personalized feeds. The platform said on March 31 that these pages suggest content based on a post’s engagement metrics and the user’s previous activity.

How X facilitates the spread of misinformation

In fact, after NewsGuard analysts repeatedly viewed the profiles of the 12 state media outlets they monitored for their research, “their accounts began to appear in sections of the platform where they were previously absent, such as the ‘Who to follow’ section or among the search results. For example, when a NewsGuard analyst created a new account “It appears that nine of these 12 state accounts – we read further – are also promoted by X’s premium subscription service which, according to the platform, allows verified accounts to receive ‘priority position in conversations’ and appear in personalized feeds” .

This premium level, experts explain, also includes the assignment of the blue check. Musk said non-premium subscribers can no longer appear in the platform’s recommendations, explaining that this is “the only realistic way to prevent swarms of AI bots from taking over.” NewsGuard contacted the 12 monitored state media outlets via email to ask them whether they agreed with the assessment emerging from the research, i.e. that they had repeatedly spread disinformation, and why they chose not to spontaneously reveal on their social media profiles that they were affiliated to governments. None of the organizations responded.

And in Italy? The case of Xinhua

NewsGuard also conducted monitoring in our country, applying the same methodology to examine the engagement of the only Italian account of Xinhua (XHItalia) on X, the only one of the 12 newspapers controlled by governments to have a profile in our language. According to Meltwater data in September 2023, the total number of likes and shares increased by 61% after Musk’s changes, despite the total number of posts published only growing by 11%. In total, in the three months following X’s policy change (between April 21 and July 20, 2023), the account gained 3,790 likes and shares on a total of 903 posts, according to Meltwater. In the 90 days before the policy change (between January 21 and April 20, 2023), the account had gotten 2,350 likes and shares on 812 posts.