Twitter announced today that the company will start labeling tweets and accounts from official media, focusing on those from China and Russia.
For clarity: we don't let state-affiliated media accounts advertise on Twitter. We’ll also no longer include them or their Tweets in recommendations, as we continue to support a free and independent press.
More on this policy and new labels: https://t.co/BY1jTO46Zc (2/2)
— Twitter Support (@TwitterSupport) August 6, 2020
The new categorisation covers accounts in the United States, Britain, France, China and Russia – the five permanent United Nations Security Council members – but will in effect only apply to outlets from China and Russia, given the high degree of control that government bodies in those countries exert on editorial decisions through funding and other forms of pressure.
The new labels, which appear on both the homepages of targeted accounts and above every tweet posted by them, have also been added to editors and prominent reporters at the outlets.
[…] Under the new rules, publicly funded outlets that have firewalls between financial revenue and editorial decisions will not be targeted, Twitter said, pointing to the British Broadcasting Corporation (BBC) and National Public Radio (NPR) as examples.
[…] As of Thursday, the company had placed the state-affiliated media label on a large number of Chinese news accounts, including those not considered government mouthpiece outlets and at least one – Caixin Global – that is not government funded.
[…] Under the rules, accounts that are labelled as state-affiliated media and their tweets will no longer be amplified through Twitter’s recommendation systems, including its home timeline, notifications and search functions, the company said. [Source]
Several Chinese media accounts, and those of their editors, had the new label on Thursday, though the precise criteria for selection was unclear to many observers:
Sixthtone too. Guess ownership is the criteria. pic.twitter.com/uOThu6VidM
— The Dude (@the__dude98) August 6, 2020
genuinely misleading. even if it operates under constraints and there are government sponsors in the background, no way caixin should be lumped with xinhua and global times etc. https://t.co/1vbVZvpz2G
— Sam Geall (@samgeall) August 6, 2020
Honestly, not that hard a call. Caixin is 30% owned by China Media Capital, which is run by Li Ruigang. Li is also Caixin's chairman. CMC was set up in 2009 with support from China Development Bank (state owned), Dongfang Huijin (state owned) and CMCDI (state owned). https://t.co/1vnmxQ724v pic.twitter.com/4k7mijSlUs
— Matt Schrader (@MattSchrader_DC) August 6, 2020
"affiliate", from a more legal perspective, emphasizes direct or indirect control. So labelling Caixin as a state affiliated media is very accurate.
— Shawn Zhang (@shawnwzhang) August 6, 2020
— Owen Churchill (@owenschurchill) August 6, 2020
While Chinese official media and diplomats have been increasingly using Twitter and other social media—many of which are banned in China—to aggressively promote the government’s views on various issues, others have also used less official and more discreet channels to support and amplify official propaganda and disinformation. In June, Twitter removed 170,000 accounts linked to the government for spreading disinformation about COVID-19, Hong Kong protests, and other issues. More recently, a report by Raymond Serrato and Bret Schafer for The Institute for Strategic Dialogue and Alliance for Securing Democracy found that a coordinated network of Twitter accounts used replies to amplify and support government messages. From the report:
The report highlights the importance of transparency and caution in assessing inauthenticity online. ISD used seven analytical signals to build a picture of a likely coordinated, inauthentic network promoting CCP propaganda and attacking anti-CCP spokespeople or institutions. Activity rates are often relied on as the way to assess possible inauthenticity of accounts or networks, but this approach misunderstands the complexity of disinformation actors and authentic human behaviour online. ISD has used a multilayered research approach to provide a much more comprehensive picture of what coordinated networks can look like in 2020 and how difficult it can be to sift them out from authentic online activity. It is hoped this combined methodology can contribute to the burgeoning field of information operations research. This research was combined with the insights from ASD’s monitoring of official CCP social media networks to better understand the parallels in overt state activity and pro-state covert networks online.
ISD analysed the data from 828,646 followers of two CCP officials and has identified a network of suspicious Twitter accounts engaged in propagating pro-CCP narratives. A large number of accounts were created in March 2020 and are characterised by several indicators of inauthenticity and coordinated messaging on issues that include COVID-19 conspiracies, Taiwan independence, the Hong Kong demonstrations, and the recent antiracist protests in the US.
[…] Of the 445,570 tweets that ISD collected, researchers found that replies made up nearly 65% of tweets, with the largest number of replies coming after President Trump’s use of the phrase, “China Virus”. The number of replies, retweets, and original tweets decreased thereafter, but replies continued to make up a larger proportion of the tweet types.
Replies primarily targeted President Trump, American diplomatic officials, Taiwan’s president, and major media such as the Voice of America, the BBC, Deutsche Welle, the New York Times, and others. The accounts also replied to the official accounts of the CCP and Chinese state-controlled media, largely with praise or in agreement with statements they had made regarding COVID-19 or rebuffs to President Trump. [Source]
At CNN Business, Donie O’Sullivan further reports on Serrato and Schafer’s research:
Accounts with usernames like @Dotard36639316, @alex75471129, and @susanblunt2020, all only set up in March of this year, tweeted at President Trump hundreds of times, often falsely claiming that America, not China, was the source of Covid-19. These accounts were among thousands identified by the researchers. Their report said the accounts were using Twitter to push talking points in favor of CCP, China’s ruling party.
In response, Twitter questioned the report and said it has seen an uptick in use from real Chinese-speaking users since the outbreak of Covid-19. Although the government in Beijing has been shown to be running fake accounts on the social network, it’s also very possible many of the accounts spamming the president with pro-CCP messaging are real CCP supporters and sympathizers.
The accounts’ activity shows how the propaganda battle over who is responsible for the outbreak of Covid-19 is now playing out in the replies to tweets from lawmakers. The research also underscores how challenging it can be to determine who is behind what are often faceless accounts online — a government, a misinformation group for hire, a concerned citizen, or someone with a different motive. [Source]
Also on Thursday, YouTube announced that it had closed 2,500 accounts linked to China as part of their effort to weed out disinformation. From Lily Kuo at The Guardian:
The Alphabet-owned company said the channels were removed between April and June “as part of our ongoing investigation into coordinated influence operations linked to China.“
The channels generally posted “spammy, non-political content,” but a small subset touched on politics, the company said in a quarterly bulletin on disinformation operations.
Google did not identify the specific channels and provided few other details, except to link the videos to similar activity spotted by Twitter and to a disinformation campaign identified in April by social media analytics company Graphika. [Source]
See the Graphika report, “Return of the (Spamouflage) Dragon.”