People attend a rally to show support for jailed former governor Basuki Tjahaja Purnama - known by his nickname Ahok. Photograph: Bay Ismoyo/AFP/Getty Images via The Guardian |
Monday, 23 Jul 2018
To pass them off as real, Alex would enliven his fake accounts with dashes of humanity. Mixed up among the stream of political posts, his avatars – mostly pretty young Indonesian women – would bemoan their broken hearts and post pictures of their breakfasts.
But these fake accounts were not for fun; Alex and his team were told it was "war".
"When you're at war you use anything available to attack the opponent," says Alex from a cafe in central Jakarta, "but sometimes I felt disgusted with myself."
For several months in 2017 Alex, whose name has been changed, alleges he was one of more than 20 people inside a secretive cyber army that pumped out messages from fake social media accounts to support then Jakarta governor Basuki Tjahaja Purnama, known as "Ahok", as he fought for re-election.
"They told us you should have five Facebook accounts, five Twitter accounts and one Instagram," he told the Guardian. "And they told us to keep it secret. They said it was 'war time' and we had to guard the battleground and not tell anyone about where we worked."
The Jakarta election – which saw the incumbent Ahok, a Chinese Christian, compete against the former president's son Agus Yudhoyono, and the former education minister, Anies Baswedan – churned up ugly religious and racial divisions. It culminated in mass Islamic rallies and allegations that religion was being used for political gain. Demonstrators called for Ahok to be jailed on contentious blasphemy charges.
The rallies were heavily promoted by an opaque online movement known as the Muslim Cyber Army, or the MCA, which employed hundreds of fake and anonymous accounts to spread racist and hardline Islamic content designed to turn Muslim voters against Ahok.
Alex says his team was employed to counter the deluge of anti-Ahok sentiment, including hashtags that critiqued opposition candidates, or ridiculed their Islamic allies.
Alex's team, comprising Ahok supporters and university students lured by the lucrative pay of about $280 (£212) a month, was allegedly employed in a "luxury house" in Menteng, central Jakarta. They were each told to post 60 to 120 times a day on their fake Twitter accounts, and a few times each day on Facebook.
'Special forces'
In Indonesia – which ranks among the top five users of Twitter and Facebook globally – they are what are known as a "buzzer teams" – groups which amplify messages and creates a "buzz" on social networks. While not all buzzer teams use fake accounts, some do.
Alex says his team of 20 people, each with 11 social media accounts, would generate up to 2,400 posts on Twitter a day.
The operation is said to have been coordinated through a WhatsApp group called Pasukan Khusus, meaning "special forces" in Indonesian, which Alex estimates consisted of about 80 members. The team was fed content and daily hashtags to promote.
"They didn't want the accounts to be anonymous so they asked us to take photos for the profiles, so we took them from Google, or sometimes we used pictures from our friends, or photos from Facebook or WhatsApp groups," says Alex. "They also encouraged us to use accounts of beautiful women to draw attention to the material; many accounts were like that."
On Facebook they even made a few accounts using profile pictures of famous foreign actresses, who inexplicably appeared to be die-hard Ahok fans.
The cyber team was allegedly told it was "only safe" to post from the Menteng residence, where they operated from several rooms.
"The first room was for the positive content, where they spread positive content about Ahok. The second room was for negative content, spreading negative content and hate speech about the opposition," says Alex, who says he chose the positive room.
Many of the accounts had just a few hundred followers, but by getting their hashtags trending, often on a daily basis, they artificially increased their visibility on the platform. By manipulating Twitter they influenced real users and the Indonesian media, which often refers to trending hashtags as barometers of the national mood.
Pradipa Rasidi, who at the time worked for the youth wing of Transparency International in Indonesia, noticed the phenomenon when he was researching social media during the election.
"At first glance they appear normal but then they mostly only tweet about politics," he said.
Rasidi interviewed two different Ahok buzzers, who detailed using fake accounts in the same fashion as that described by Alex. Both declined to speak to the Guardian.
A social media strategist who worked one of Ahok's opponents campaigns said buzzing was a big industry.
"Some people with influential accounts get paid about 20m rupiah ($1,400/£1,069) just for one tweet. Or if you want to get a topic trending for a few hours, that costs between 1-4m rupiah," Andi, who only wanted to be identified by his first name, explained.
Based on its study of the buzzer industry in Indonesia, researchers from the Center for Innovation and Policy Research (CIPG) say all candidates in the 2017 Jakarta election used buzzer teams – and at least one of Ahok's opponents skilfully created "hundreds of bots" connected to supporting web portals.
The Baswedan campaign denied using fake accounts or bots. A Yudhoyono spokesman said they did not breach campaigning rules.
Slander, hatred and hoax
The authorities have made moves to crack down on fake news and the spread of hate speech online but buzzers, which operate in a grey area have largely slipped through the cracks.
Even the central government appears to employ such tactics. The Twitter account @IasMardiyah, for example, which Alex says was utilised by his pro-Ahok buzzer team, now posts a steady flow of government messages and propaganda for President Joko Widodo – mostly retweets about Indonesia's infrastructure and diplomatic successes, or the need to protect national unity.
Featuring an avatar of a young woman wearing a headscarf and sunglasses, the account tweets almost exclusively pro-government content with accompanying hashtags.
Recently the account has posted about Indonesia's election to the United Nations security council, fighting terrorism, boosting agricultural exports, a new airport in West Java, next month's Asian Games, but also on sensitive issues such as West Papua.
A presidential spokesperson was asked for comment by the Guardian, but did not respond.
A spokesperson from Twitter declined to specify how many fake Indonesian accounts it had identified or removed from its platform in the past year. The company said it had "developed new techniques and proprietary machine learning for identifying malicious automation".
The Baswedan campaign denied using fake accounts or bots. A Yudhoyono spokesman said they did not breach campaigning rules.
Slander, hatred and hoax
The authorities have made moves to crack down on fake news and the spread of hate speech online but buzzers, which operate in a grey area have largely slipped through the cracks.
Even the central government appears to employ such tactics. The Twitter account @IasMardiyah, for example, which Alex says was utilised by his pro-Ahok buzzer team, now posts a steady flow of government messages and propaganda for President Joko Widodo – mostly retweets about Indonesia's infrastructure and diplomatic successes, or the need to protect national unity.
Featuring an avatar of a young woman wearing a headscarf and sunglasses, the account tweets almost exclusively pro-government content with accompanying hashtags.
Recently the account has posted about Indonesia's election to the United Nations security council, fighting terrorism, boosting agricultural exports, a new airport in West Java, next month's Asian Games, but also on sensitive issues such as West Papua.
A presidential spokesperson was asked for comment by the Guardian, but did not respond.
A spokesperson from Twitter declined to specify how many fake Indonesian accounts it had identified or removed from its platform in the past year. The company said it had "developed new techniques and proprietary machine learning for identifying malicious automation".
0 comments:
Post a Comment