Twitter puts its foot down on fake followers
Twitter has finally taken action to suspend thousands of fake accounts in South and Southeast Asia, a problem that has ballooned over the last four months during which Cambodia, Malaysia and Pakistan held national elections.
Twitter suspended the accounts earlier in August, after months of complaints about them spreading fake news that in some cases had led to violence. An analysis by the Australia-based programmer Nisal Periyapperuma showed that Twitter suspended roughly 6,000 fake accounts, or automated accounts now commonly referred to as bots, in Cambodia alone. A Twitter spokesperson told the Nikkei
Asian Review in late July that the company had taken steps to delete fake accounts in the region but that it also had a duty to ensure free speech: “We are committed to providing a service that fosters and facilitates free and open democratic debate and that promotes positive change in the world.”
But Twitter is facing a backlash across the world from both left and right of the political spectrum as it does not believe it should police content and as it sees itself as simply a platform. In late July, US President Donald Trump lashed out at the company, accusing it of “shadow-banning” Republican politicians, or hiding content from some users without informing them.
Trump t weeted then: “Twitter ‘SHADOW BANNING’ prominent Republicans. Not good. We will look into this discriminatory and illegal practice at once! Many complaints.”
In Asia, there are mounting worries that bots are influencing the way people vote, as they were believed to have done in the US presidential election in November 2016 and the UK vote to leave the European Union in June the same year.
While the Berlin-based digital researcher Raymond Serrato said the majority of Asian accounts are still automated spam accounts or commercial bots, others are increasingly sharing political content. For example, in Myanmar, some of these accounts send out anti-Rohingya messages.
“In some few cases, accounts that were created at about the same time — and shared characteristics, like names with sequences of digits — were sharing anti-Rohingya and pro-military content,” Serrato said.
Those tweets blamed Rohingya in general for attacks on the police and spread anti-Muslim content. Some believe these accounts to be operated by the military, although this has not been substantiated. Serrato said the frequency of the tweets suggested they were sent by humans, but their source remains unclear.
In Malaysia, an analysis by the Digital Forensic Research Lab of the Washington-based Atlantic Council showed that bots started spreading anti-opposition content about three weeks before the elections in early May.
Bots shared anti-opposition hashtags “#SayNOtoPH and #KalahkanPakatan — referring to the opposition coalition — thousands of times from April 12. According to the study, 98.4% of the 22,000 accounts that used the hashtag had a “low authority score”, which indicates that they were automated accounts.
In Cambodia, a similar trend emerged at election time in July. Analysis shows that the accounts that tweeted political content were all created around the same time. Dozens suddenly and uniformly displayed profile photos within seconds of following other accounts, as if to prove their existence. Many of them also had Facebook accounts created on the same day under the same name, posting political propaganda.
These accounts target “influential” Twitter users, such as journalists, actors and researchers. Two weeks before Cambodia’s elections on July 29, bots flooded Twitter with propaganda backing the ruling Cambodian People’s Party, hijacking the hashtags #Electionkh and #electionskh that were previously used by journalists and others to stay informed.
Some well-known journalists in Thailand also reported a spike in the number of followers on their Twitter accounts earlier this year. There was speculation that groups attempting to influence opinion leading up to the country’s next general election — if and when it happens — could have been responsible.
Fake profiles also posted pro-CPP propaganda under another hashtag previously used by the opposition to call on people to abstain from voting: #Cleanfinger.
In Sri Lanka, fake accounts were blamed for spreading anti-Muslim messages. Early in March, the government banned social media after race riots, in part fuelled by fake news reports, left several dead.
Twitter said it would continue to strengthen its mechanism against manipulation. “Twitter deeply respects the integrity of the election process, which is a cornerstone for all democracies. We continue to strengthen Twitter against attempted manipulation, including malicious automated accounts and spam, as well as other activities that violate our terms of service,” the spokesperson said.
But deleting bots alone does not fully address the problem, Serrato said. Hundreds of new ones are created every day, suggesting that the bodies behind these accounts are unlikely to stop soon.
“I think we’re likely to see more sophisticated ways of manipulating people online because there is a market for this kind of work,” he said.
WhatsApp, owned by Facebook, is another platform through which fake news has been spread in Asia. In India, there have been cases of vigilante killings over the last year after false rumours were spread about child abductions on WhatsApp, highlighting the severity of the problem of fake news in the region.