With the 2020 US Election coming up, and candidates in disagreement over how social platforms should handle political speech, this new disclosure from Twitter comes at an interesting time.
Today, Twitter has outlined its latest account removals – amounting to more than 32,000 accounts – which had been found to be participating in various coordinated manipulation campaigns.
As explained by Twitter:
“Today we are disclosing 32,242 accounts, including three distinct operations that we have attributed to the People’s Republic of China (PRC), Russia, and Turkey respectively. Every account and piece of content associated with these operations has been permanently removed from the service. In addition, we have shared relevant data from this disclosure with two leading research partners: Australian Strategic Policy Institute (ASPI) and Stanford Internet Observatory (SIO).”
The networks each had their own purpose in amplifying different content.
- The Chinese network of 23,750 accounts were “Tweeting predominantly in Chinese languages and spreading geopolitical narratives favorable to the Communist Party of China (CCP), while continuing to push deceptive narratives about the political dynamics in Hong Kong”. Twitter says that it additionally identified a network of more than 150,000 “amplifier accounts” which were designed to boost the content from this core network.
- The Russian network of 1,152 accounts was linked to state-backed media website ‘Current Policy’, and was found to be cross-posting and amplifying content “in an inauthentic, coordinated manner for political ends”. The main aim of this cluster was to promote the pro-Kremlin United Russia party, and to attack political dissidents.
- The Turkish group of 7,340 accounts was found to be amplifying support for the AK Parti and President Recep Tayyip Erdogan. “Technical signals point to the network being associated with the youth wing of the party and a centralized network that maintained a significant number of compromised accounts.”
Of the three, the Chinese network seems the most concerning – and in particular, the expanded network of “amplifier accounts” which were designed to engage with tweets from the main network in order to increase their reach and ranking in feeds. Why Twitter has seemingly not provided full disclosure on this expanded network is unclear, but the fact that massive clusters of 170k accounts are being formulated for such purpose – even if, in this case, Twitter says they did not end up gaining significant traction – is a major concern.
Earlier in the year, researchers from Queensland University of Technology in Australia identified what they believed to be a huge cluster of bot accounts which were amplifying disinformation about the cause of the bushfires that hit the east coast of the nation. In that instance, the bot accounts were apparently seeking to boost discussion of the fires being started by arsonists, in order to drown out the debate around climate change that was gaining traction at the time.
It’s not clear who was behind that campaign, or if indeed it was a coordinated bot effort, but the finding aligns with other discoveries in the past, where huge armies of bot accounts have been used to clog the network with a specific political leaning.
For its part, Twitter said in March that the identification of bot accounts by third parties is largely flawed, and mostly incorrect, so reports like these can’t necessarily be trusted. Twitter has also said that the amount of fake accounts on its platform is less than 5% – though it theoretically wouldn’t count bot accounts as fake, because some bot accounts are helpful and shouldn’t be seen as fake, as such.
But this latest disclosure suggests that bot swarms like this are indeed possible, and are being used for political manipulation.
With the Chinese network in particular, Twitter says “they were largely caught early and failed to achieve considerable traction on the service”. So while the network was large, Twitter’s systems picked it up quickly, which highlights the benefits of its evolving approach. Which is good, but it does bring into question whether other networks like this, that have been identified in the past, are also still active, and causing an impact.
Bot networks were particularly present, based on reports, during the 2016 US Presidential Election. In one example, researchers uncovered “huge, inter-connected Twitter bot networks” seeking to influence political discussion, with the largest incorporating some 500,000 accounts. Last year, Wired reported that bot profiles were still dominating political news streams, with bot profiles contributing up to 60% of tweet activity around some events.
On one hand, today’s disclosure gives me more faith that Twitter’s getting better at detecting such activity, yet on the other, the size of the Chinese amplifier network in particular lends more weight to past, similar claims, which Twitter has at least partially denied.
And while people will point to the fact that Twitter has fewer users than other networks, limiting any related impact either way, the fact of the matter is that Twitter is a critically influential network. You might not even use Twitter yourself, but the news that you see will often come from tweets, with highly engaged influencers, in various forms, locked into the real-time tweet stream, from which they formulate and share opinions across to other platforms.
The fact that state-backed groups are focused on influencing the Twitter conversation should be indicative enough – they don’t undertake these activities because Twitter isn’t valuable in this respect.
If you want to influence opinion, Twitter is now where many start. Which is why it’s increasingly important that Twitter seeks to improve its detection and removal processes wherever it can.