Taipei, Taiwan – Scrolling through Twitter during the past few weeks, Sarah Hurst, an independent journalist in the United Kingdom, began noticing changes in the way the platform displayed certain government and media accounts.
Hurst, who writes about Russia and Ukraine, started seeing more tweets from Russian government accounts, Russian state media and government mouthpieces in her “For You” tab.
Twitter launched the tab in January as the new default view for users, displaying algorithmically-selected tweets along with those from accounts the user follows.
“Previously on your settings, you could choose to have tweets organised by ‘top tweets’ or chronologically in your newsfeed. I usually had it on ‘top tweets’ so I didn’t miss the biggest news stories,” Hurst told Al Jazeera.
“Now, whenever I open ‘For You,’ I see a stream of Russian and Chinese government propaganda accounts.”
“I have been campaigning against [Russian President Vladimir] Putin since he annexed Crimea in 2014,” Hurst added.
Whether by accident or design, Twitter is giving greater prominence to government and state-run media accounts that are often criticised for spreading disinformation, such as Russia’s RT and China’s Global Times.
The changes appear to include subtle boosts to state-affiliated accounts as well as more prominent placement in users’ feeds.
In a Substack post last month, Wenhao Ma, a reporter for the United States government-funded Voice of America, said the top search results for the name of Taiwanese President Tsai Ing-wen and the phrase “US-China” were posts from Chinese state media denouncing Tsai and the US government.
The latest changes at the social media giant, which has undergone significant upheaval since its takeover by billionaire tech leader Elon Musk, have prompted concern among journalists and disinformation scholars – as well as broader debate about what counts as disinformation or propaganda and who gets to decide.
Some analysts argue the apparent changes to Twitter’s algorithm have troubling implications for the public discourse, democracy and the future of the platform itself.
“It’s clear the algorithm has changed and what counts as being ‘for you’ has changed,” Darren Linvill, an associate professor at South Carolina’s Clemson University who researches social media disinformation, told Al Jazeera.
“Historically, Twitter has been really good at giving people only stuff that they’re vaguely looking for, so I think a normal person would be surprised that most of Twitter is some combination of K-Pop and porn. And it didn’t give you that. Elon is trying to shake things up and clearly, he’s messing with an algorithm that they’ve developed over more than a decade to be successful in giving people what they want.”
Timothy Graham, a senior lecturer in digital media at the Queensland University of Technology (QUT), said the site’s recommendations were now “a bit out of control” compared with the more curated approach of before.
“Some people are seeing harmful war propaganda relating to the Ukraine war coming from Russia and ministers, or from diplomatic accounts, or from Russia Today,” Graham said.
Prior to Musk’s purchase of Twitter last October, the platform had taken steps to reduce the reach of certain state-affiliated accounts.
In 2020, Twitter introduced the label “state-affiliated media”, which it defines as “outlets where the state exercises control over editorial content through financial resources, direct or indirect political pressures, and/or control over production and distribution.”
In practice, this label was almost exclusively applied to Russian and Chinese government media accounts, although Twitter originally said the label would be rolled out to media from the five countries that sit on the United Nations Security Council – China, Russia, France, the United States and the UK.
The platform’s guidelines ultimately excluded outlets like the BBC in the UK and National Public Radio, also known as NPR, in the US, both of which receive government funding but are widely regarded as editorially independent.
Those rules, however, are changing – sometimes on a practically daily basis.
Last week, Musk appeared to change Twitter’s definition of “state-affiliated media” by briefly adding the label to NPR, an outlet that right-leaning Americans often accuse of having a liberal bias.
The label was removed within days after blowback from critics who defended NPR’s record of editorial independence and noted that government funding accounts for only 2 percent of the outlet’s budget, although some US conservatives and Chinese-state media employees welcomed the state media tag.
Twitter added a new “government-funded media” label to NPR’s account in its place.
The designation, which has also been applied to the BBC, PBS and Voice of America in recent days, refers to “outlets where the government provides some or all of the outlet’s funding and may have varying degrees of government involvement over editorial content”.
The new label has not been added to some other state-funded media, including Al Jazeera and France 24, which are funded by the Qatari and French governments, respectively.
Twitter did not respond to a request for comment.
But in an interview with the BBC on Wednesday, Musk said: “We want it as truthful and accurate as possible – we’re adjusting the label to “[the BBC being] publicly funded – we’ll try to be accurate.”
Earlier this week, Musk was quoted telling one of NPR’s reporters: “If you really think that the government has no influence on the entity they’re funding then you’ve been marinating in the Kool-Aid for too long.”
Taken together, the changes at Twitter make it easier to spread propaganda and “fake news” about current events, including major conflicts like the war in Ukraine, said Graham, the QUT lecturer.
“Some of them are conspiracy theories about neo-Nazis taking over Ukraine. All of them are trying to justify and provide this narrative justification for what Russia is doing, and all of them are trying to appeal to audiences who will amplify them in one way or another,” he said.
“It’s not this kind of hypodermic needle model as if they inject you with false information, and then you believe it and start spreading it,” Graham added.
“But it’s more than like we get into a situation where these false and misleading narratives that pollute and just cause chaos …[and] get oxygenated by the rest of the media ecosystem. When that happens, it’s payday, essentially, for state media.”