Twitter kills thousands of misinformation accounts

Social media platform Twitter has suspended 32,242 state-linked misinformation accounts from China, Russia and Turkey, and published their details to a public archive of misinformation resources that it runs alongside the Australian Strategic Policy Institute (ASPI) and the Stanford Internet Observatory (SIO).

The accounts were all suspended for violating the service’s platform manipulation policies, and in making a full public disclosure, Twitter said it hoped to “serve the public conversation, remove bad actors and … advance public understanding of these critical topics”.

Most of the removed accounts – 23,750 – plus a further 150,000 “amplifier” accounts that boosted their content, were linked to the Chinese state, said Twitter, which first identified suspicious activity from the network in August 2019.

The core accounts were largely caught early and never achieved much traction, with low follower counts and engagement, while the amplifier accounts were designed to artificially boost impression metrics on the service by engaging with the core network through likes, replies and retweets.

The Chinese network was nevertheless quite dangerous and was involved in a range of manipulative and coordinated activities designed to improve public perception of the Chinese government and spread misinformation about political movements in Hong Kong and Taiwan, the Covid-19 coronavirus outbreak, and exiled businessman Guo Wengui, who fled China after becoming embroiled in a corruption scandal. It mostly targeted Twitter users in Hong Kong and the overseas Chinese diaspora.

The Russian network of 1,152 accounts was largely associated with a state-backed propaganda media website, Current Policy, and were suspended for cross-posting and amplifying content favourable to Russian president Vladimir Putin in an inauthentic manner, including boosting his United Russia political party and attacking dissidents.

The Turkish network of 7,340 accounts was detected early in 2020, and largely concerned itself with domestic audiences within Turkey, amplifying political messages favourable to president Recep Tayyip Erdoğan.

Twitter and its partners assessed that the network was associated with the youth wing of the AK Parti, Erdoğan’s political party, which maintained a centralised network consisting of many compromised accounts.

As a result of this, some of the accounts listed in Twitter’s archive are associated with organisations critical of the Turkish government, which has been engaged in a systematic roll-back of the country’s multi-party, liberal democracy since 2013. The network was also used for cryptocurrency spam.

Twitter said its misinformation archive and site integrity efforts represented some of the most critical work the organisation does “to protect the public conversation”.

Moving forward, Twitter highlighted three steps it plans to take to make this work more transparent. It said it would: offer more clarity in its archive around impression counts and try to better measure tangible impacts of misinformation operations; formalise its academic partnerships to ensure they are globally diverse and effective at enhancing public understanding of misinformation; and facilitate future discussions between experts, industry and governments.