The volume of COVID misinformation has jumped alarmingly on Twitter, while anti-vaccination networks that had formerly been dispersed are already reforming and reorganising, according to research.
Key points:
When Elon Musk took over the social media platform in late October, he set about dismantling certain moderation policies as well as the teams tasked with enforcing them.
By the end of November, a system that had been painstakingly built by many people over many years lay in tatters.
Mr Musk made a special point of allowing COVID misinformation, announcing on November 30 that his new company was "no longer enforcing the COVID-19 misleading information policy".
Given these changes, researchers expected to see an increase in COVID misinformation when they analysed the volume of tweets containing certain keywords associated with established conspiratorial tropes, such as "Covid AND hoax", "Fauci AND lied" and "Ivermectin".
Now that data is in — and the size and speed of the increase has surprised even these seasoned observers.
"This is a watershed moment," said Timothy Graham, the misinformation expert at the Queensland University of Technology who ran the analysis.
"This is a clear signal that COVID misinformation is back on the menu."
A new documentary drove the return of COVID misinfo
The graph below shows the hourly volume of tweets that include at least one of seven COVID misinformation keywords, or keyword combinations, published from November 1 to December 4.
This graph plots the number of COVID misinformation tweets from November 1 to December 4.
The spike in the second half of November is partly due to the launch of an anti-vax propaganda documentary, Died Suddenly, which spread widely and served as a timely vehicle for the return of COVID misinformation on Mr Musk's Twitter.
Whereas once it would probably have been banned, the Died Suddenly Twitter account has amassed over 155,000 followers and features the $8-per-month blue checkmark, which means the owner has paid for a premium service that will promote their tweets.
"I think that it would have been specially moderated if the Twitter safety team were still working at Twitter," Dr Graham said.
The spike wasn't entirely due to the appearance of a new documentary: There were also a lot of tweets about Ivermectin, a drug that anti-vaxxers have been promoting for more than a year as a "miracle" COVID cure.
Other misinformation terms are also being tweeted more often as well.
The number of COVID misinformation tweets from November 1 to December 4 is shown in this graph, minus those that refer to the Died Suddenly documentary or ivermectin.
"That little collection [of terms] gives us a window into the related conspiracy theories that are getting whipped up and getting recharged with new energy and freedom to circulate," Dr Graham said.
"As Twitter is allowing … tweets about Ivermectin and this debunked documentary, business is booming again for all these other ones."
No evidence tweets are being demoted
Last week, research conducted by the Centre for Countering Digital Hate found the volume of hate speech on Twitter had grown dramatically over the past month.
The use of the n-word on Twitter is triple the 2022 average and the use of slurs against gay men and trans persons is up 58 per cent and 62 per cent, respectively.
In response, Mr Musk said "hate speech impressions" were lower.
The volume of tweets may have gone up, but moderators were restricting their reach, so these tweets were being seen by fewer people, Mr Musk claimed. This moderation practice is known as "shadow banning".
Mr Musk's claim can't be verified.
"This metric of reach is not available to researchers and journalists," Dr Graham said.
"[Using this metric] gives them the power to be able to make claims about stuff like this, and to not be accountable."
That said, it is possible to indirectly measure reach through the metrics of engagement, such as the number of likes, replies or retweets each tweet receives.
So, Dr Graham checked these figures for the COVID misinformation tweets he had already gathered.
There was no evidence of shadow banning of COVID or vaccine misinformation.
The four graphs show (respectively, from top), the number of likes, retweets, replies and quote tweets each original COVID misinformation tweet received, from November 1 to December 4.
"If they were shadow banning users … we would see the tweet volume increasing, but we'd see the engagement metrics going down," Dr Graham said.
"But it's all going up."
Disbanded anti-vax networks reforming
Is the misinformation being spread by lone actors or coordinated networks?
The chart below shows the connections between each account in Dr Graham's dataset. Accounts appear as nodes, with lines of connection indicating two accounts have tweeted the exact same tweet at least twice.
If there was no coordination, this chart would be full of disconnected dots.
"The graph suggests they are coordinating. There is a strategy here to leverage Twitter, to build something," Dr Graham said.
"There's really a groundswell of coordinated mobilisation."
A common theme of these tweets is the belief that "suddenly this is a new age of free speech under Elon Musk".
The spread of COVID misinformation may ultimately cost lives and undermine trust in medical expertise, but it's also a sign of what's to come.
For years, Twitter has served a vital function as an information-sharing and verification service. That's being very rapidly eroded.
The events of the past month have shown not only how central Twitter is within the "global media ecology", but its fragility — how easily an essential service used by millions of people around the world can be dismantled at the whim of the world's richest man, Dr Graham said.
"Twitter is basic infrastructure now — it's like a road, it's like rates and rubbish, we depend on it," he said.
"It's the worst, it really is the worst; we're back to a situation where medical expertise at the highest level … is being undermined."
