An anti-China conspiracy theory is being boosted on social media by a large network of automated accounts that also promote messages supporting U.S. president Donald Trump, research has found.

The suggestion that COVID-19 is a bioweapon - widely dismissed by experts from the U.S. intelligence agencies, the World Health Organization and politicians and scientists in the UK, Europe and elsewhere - has appeared frequently on Twitter.

Researchers from the Queensland University of Technology, commissioned by the Australia Institute's Centre for Responsible Technology, over a ten day period in late March examined 2.6 million tweets and 25.5 million retweets related to COVID-19. 

To distinguish authentic Twitter users from bot-like accounts the team used a combination of reading by humans with a model that accounted for around 1,000 variables.

"If you've got two accounts who are constantly retweeting the same content, exactly the same tweet over and over and over again within one second of each other, so either exactly at the same time or within a second, then there's a very, very low probability that this has happened by chance for two organic individuals," explained Timothy Graham, who along with Alex Burns carried out the study. 

Using this methodology the team uncovered a network of 5,752 accounts that, in a coordinated fashion, retweeted COVID-related content 6,559 times.  

The analysis further identified a network of 2,903 accounts and 4,125 links or co-retweets between them, which promoted the conspiracy theory that COVID-19 is a bioweapon developed by China. The majority of the clusters within this network identify as Pro-Trump, QAnon, and/or Republican partisan.

READ MORE: Non-white Britons more at risk of COVID-19 death, report reveals

The QAnon conspiracy theory pushes a narrative centered on a secret war between the U.S. 'deep state' and Trump. 

The study focussed on 882 original tweets that focussed on the bio-weapon conspiracy theory, which were retweeted 18,498 times and liked 31,783 times, generating an estimated five million impressions on the platform.

"What's really interesting about that is that these accounts are predominantly associated with, or their profiles identify as, pro-Trump, kind of 'Make America Great Again' accounts. These are standard fare, pretty much in disinformation analysis and in political analysis," said Graham.  

The researcher notes Trump's "unsubstantiated and vague claims" on the origins of COVID-19. "This stuff has pretty much been picked up on by these bot-like accounts and more so I think by trolls, by sock puppet troll-like, human-controlled accounts...it's a feedback loop. 

"As soon as it's validated by an elected official or by a celebrity or by mainstream media like Fox News, for example, it's on for all money, essentially. And you see them just posting this over and over again."

Identifying bot networks opens up the debate around the responsibility of social media companies to monitor their platforms for coordinated misinformation. Earlier this week Twitter, in an unprecedented move, censored a Trump tweet, threatening the Black Lives Matter protesters, for "glorifying violence." While Facebook allowed the same post on it's platform without a label. 

Since the early months of the pandemic, both platforms have engaged in various measures to stop the spread of misleading or false news around COVID-19. Although both the institute's research and a recent report by the Center for Countering Digital Hate (CCDH), which noted hundreds of posts spreading misinformation about COVID-19 being left on social media, suggests the problem remains. 

Responding to the report from the UK-Based CCDH, which detailed posts left up relating to false cures and conspiracy theories around 5G, a Twitter spokesperson said: "We're prioritizing the removal of COVID-19 content when it has a call to action that could potentially cause harm."

"As we've said previously, we will not take enforcement action on every tweet that contains incomplete or disputed information about COVID-19. Since introducing these new policies on March 18 and as we've doubled down on tech, our automated systems have challenged more than 4.3 million accounts which were targeting discussions around COVID-19 with spammy or manipulative behaviors."

READ MORE: Sweden's top COVID-19 scientist backtracks on country's response

Despite this, the institute's research suggests a significant amount of influential misinformation is still proliferating on social media. "I'm surprised that Twitter and Facebook, for example, aren't able to stop these kinds of coordinated amplification clusters before they do too much damage," notes Graham. 

The Australian report's authors recommend that platform operators improve detection and mitigation of bot activity - especially on a claim which has such a potentially serious impact and has been so comprehensively dismissed by the scientific community. 

"If you feel that people have a right to be able to say whatever their opinion is about that, then there's not much that we can do," Graham admits.

"But I think in this case, if there are lines that have been crossed, I think that the platforms really have a responsibility to root out as quickly as possible when there are coordinated attempts to push this." 

Check out The Pandemic Playbook, CGTN Europe's major investigation into the lessons learned from COVID-19