Twitter, Facebook continue to miss blatantly inauthentic COVID-related activity
90% of the most active accounts are still active despite posting around 1000 tweets per day in some cases. Facebook had not marked any of the conspiracy posts we reviewed either.
This report describes and observes COVID-related tweets for July 2022. The dataset came from the COVID-19 Misinfo Portal, a repository of tweets collected via Twitter's COVID-19 Streaming Endpoint. The text will specifically state the content location when the posts or accounts described are not on Twitter.
Most active accounts tweeting about COVID
A network of inauthentic accounts posting Japanese-language spam dominated the top spots for most active accounts for COVID-related tweets in July 2022. The data show that 90% of the most active accounts tweeting about COVID are still active on the platform, despite some users posting at a rate of over 1000 posts per day. The DFR Lab stated:
“The benchmark for suspicious activity varies. The Oxford Internet Institute’s Computational Propaganda team views an average of more than 50 posts a day as suspicious; this is a widely recognized and applied benchmark but may be on the low side. (The DFR Lab) views 72 tweets per day (one every ten minutes for twelve hours at a stretch) as suspicious, and over 144 tweets per day as highly suspicious.”
Twitter has either failed to detect or chosen not to address this network. That the network could have gone unnoticed is difficult to believe. Many of the tweets are identical or nearly identical.
At least nine users (discussed in more detail below) out of the top ten most active accounts appear to be fake and exclusively devoted to posting spam in Japanese-language. Eight of the nine bot-like accounts remain active on Twitter. Examples of spam posts from these accounts can be found here: 1, 2, 3.
The most active account does not attempt to portray itself as a human-operated account and appears to promote a health-related app. We did not assess whether the app was legitimate or not as it was not portraying a human, and further detail existed outside the scope of this report.
The Japanese-language spam network and each account’s status can be found below:
Status: Active Suspended (12/1/2022)
Status: Active (12/1/2022)
Status: Suspended
Status: Active Suspended (12/1/2022)
Status: Active Suspended (12/1/2022)
Status: Active Suspended (12/1/2022)
Status: Active (12/1/2022)
Status: Active Caution (12/1/2022)
Status: Active (12/1/2022)
Most shared URLs in COVID-related tweets
Among the most-shared URLs on Twitter for July 2022, we found multiple links to pages promoting a conspiracy theory about the US creating COVID-19 at Fort Detrick, a false claim promoted by the Chinese government. The number one link led to a Tumblr account that was devoted to the “Milk Tea Alliance.”
The Tumblr page featured a cartoon implying the US created COVID-19 at Fort Detrick, although the image misspells it as “Fort Derrick.”
Other media in the Tumblr post included videos hosted on Vimeo and YouTube. YouTube has a policy against COVID misinformation, although the embedded video may not cross the threshold for action. The publication date is January 10, 2022, a day after the user joined, and it has 1,821 views. The video description reads:
The COVID-19 come from Fort Detrick and they've been trying to cover up the truth. That's terrible!
The awkward grammar may indicate the use of translation software or that the author was not a native English speaker.
The “Milk Tea Alliance” on Zenodo
The Tumblr page also linked to a Zenodo report that appears superficially to be a scientific study. The “Milk Tea Alliance” report first appeared on the unrestricted pre-print website Zenodo in December 2021. It was later edited in May 2022.
Another influential piece of disinformation purporting to show the virus came from a lab appeared on Zenodo in September 2020. Much like the September 2020 piece, this report is a fictional amalgamation of biological weapons disinformation and conspiracy theories. The Milk Tea Alliance report differs in that it portrays the United States as having created the coronavirus, and it goes so far as to name and accuse a prominent US scientist of involvement.
By naming real people and events in the story, the author leaves the reader with real-world data points that can be verified, potentially increasing its believability. Naming a person also creates a potential focal point for groups to direct outrage and could constitute an atypical form of targeted harassment. Intentionally, inserting specific people into a conspiracy story might be an effective means of discouraging them from speaking publicly. Twitter campaigns have been used in Africa to encourage activists to self-censor.
Although Zenodo requires no screening or review, a reader’s impression may be otherwise. As a dramatic example, one could copy and paste Taylor Swift lyrics and publish them on Zenodo. The website does have a warning, although the effectiveness of that warning is unknown.
Papers are not peer-reviewed by Zenodo and must be regarded as preliminary until peer-reviewed by multiple experts in the field. Thus it should not be regarded as conclusive or be reported in news media as established information, as the main claims may not stand the test of scientific scrutiny.
The recommended citation reads:
Milk Tea Alliance. (2022). Investigation Report on COVID-19 Transmission:v1.1.0 (v1.1.0). Zenodo. https://doi.org/10.5281/zenodo.6534088
Facebook posts for "Milk Tea Alliance"
To see if "Milk Tea Alliance" content was spreading across platforms, we searched for "milk tea alliance" on Facebook. Although the links were from Live Journal, they contained the same graphics, YouTube video, and link to the Zenodo report that we found in the Tumblr link shared on Twitter.
Multiple posts with "Milk Tea Alliance" links contained text that appeared in posts from other accounts, potentially suggesting an inauthentic network on Facebook as well. Facebook had not marked a single post with a warning or fact check.
The Other Milk Tea Alliance
The hashtag "MilkTeaAlliance" was described in a report from the Media Manipulation Casebook:
"The Milk Tea Alliance is an online coalition of youth mainly from Thailand, Hong Kong, Taiwan, and Myanmar. It originated in April 2020 after pro-Chinese Communist Party (CCP) internet accounts launched an online campaign to harass a Thai celebrity and his supporters. Soon, a loosely coordinated group of young, primarily Southeast Asian, pro-democracy netizens came together to offset the pro-CCP activities. This culminated in a meme war on Twitter between the two sides.
Once the meme war with the pro-CCP supporters ended, the online activists began to coalesce as a group to share petitions and activist-oriented campaign materials under the banner of "Milk Tea Alliance." Today, the Milk Tea Alliance is a loose international network of youth engaged in media manipulation, using the hashtag #MilkTeaAlliance, to battle what they view as authoritarianism worldwide, whether targeting the CCP or their own governments."
The evidence was insufficient to say whether the report and the Milk Tea Alliance described in Media Manipulation Casebook have any relation. Another party could have easily chosen the name in a bid to use the existing MilkTeaAlliance network as a cover to spread disinformation.
The Milk Tea Alliance described by the Casebook has used terms like "ChineseVirus" and had a moment of harmonizing with the American Republican Party, the major conservative party in the United States. Although figures from that party have speculated that the coronavirus was somehow "created in a lab," none have claimed it was inside the US.
Data source: Gruzd, A. & Mai, P. (2021). COVID-19 Misinformation Portal – A Rapid Response Project from the Social Media Lab at Toronto Metropolitan University. Retrieved from https://covid19misinfo.org/