Troll With It - How Twitter Bots Threaten Our Democracy

On the internet, nobody knows you’re a dog. So goes the adage, first coined by New Yorker cartoonist Peter Steiner as a comment on internet anonymity.

Little did Steiner know that this statement – made in 1993 – would prove so relevant in the current era of digital disinformation. Replace the word ‘dog’ with ‘bot’, and what we’ve got is an accurate depiction of the current online landscape, which is being increasingly influenced by nefarious agents.

So, what is a bot? On a very basic level, bots are automated software, used to carry out tasks on behalf of humans. Their earliest usage included sending spam emails; and gathering lots of data on a particular subject – in which guise they are known as web scrapers. Recent research from the University of Southern California and Indiana University found that there are approximately 50 million bots currently active on Twitter – that’s approximately 15% of total users. In part, this is because Twitter’s API is an open platform, which enables real-time posting and parsing of information.

Bots rose to infamy during the 2016 US Presidential election, after revelations that Russian-affiliated groups deployed a combination of social, political, and cyborg bots to infiltrate the flow of news, and to back Donald Trump’s candidacy. Remarkably, most of this covert work originated in a troll-farm in St. Petersburg.

INFLAMMATORY TEXT
There was a similar eruption of ‘bot activity’ during the hugely divisive Brexit campaign in the UK, during which, according to research carried out by Oxford Universities Computational Propaganda (COMPROP) department, the two most active accounts from both sides of the debate were bots with the handles @ivoteLeave and @ivotestay.

Here in Ireland, there are growing concerns that bots have infiltrated Irish social media platforms, particularly Twitter. Last year, broadcaster Philip Boucher Hayes spoke of how he was targeted by an army of bots, who began following him in their thousands. In January 2018, porn bots – pretend ‘sex kittens’ – infiltrated the disclosures tribunal by spamming the hashtag with ‘indecent’ content and making it impossible to search for genuine updates.
When you consider the importance of social media in shaping contemporary political debate, none of this is really surprising.

In an Irish context, the upcoming referendum on the Eighth Amendment has already spurred the deployment of bots on Irish Twitter in a concerted attempt to distort the issue. In just a few short months, active Irish Twitter users have noticed a spike in Twitter accounts claiming to be Irish citizens, who are targeting Repeal activists with inflammatory text and pictures.

“It’ll be something antagonistic like, ‘So you think killing babies is okay?’” explains Stephanie Fleming, volunteer at @repeal_shield, a new Twitter account dedicated to detecting and blocking ‘pro-life’ bot accounts. “It’s like these accounts are trying to take you on some kind of tangent. It’s rarely anything to do with the content that you are talking about. I can only speculate about what the purpose is, but when it comes to the actual astroturfing and bots, the point is to take energy.”

Anamaria Dutceac Segesten, senior lecturer of European Studies at the University of Lund, has been researching bots and their impact on political discourse. She is keen to emphasise that the bots themselves are not the problem.

“A bot isn’t necessarily something that is evil,” she notes. “It’s all about the uses and the functions. Social and political bots are a specific set of bots. They are made with different functions – sometimes they are quiet, and sometimes they are interacting with humans. The more malicious ones are what I like to call a cyborg – a hybrid human-computer mix.”

ANTI-CHOICE PROFILES
Cyborgs are probably the most accurate description of what we’re currently seeing on Irish Twitter. These accounts act as an amplifier for fringe ideas or movements that have low representation online, but want to create an impression of momentum and popularity.

One or two people can be enough to pollute the atmosphere dramatically if they use multiple networked accounts – the number sometimes reaches into the hundreds – which retweet, like and comment on each others’ profiles, whilst attacking those with opposing views.

“A cyborg finds their target by searching for keywords specific to the debate and your geo-location,” says Segesten. “They do not customise the content to your message, they’ll just repost it to other cyborgs, and the intention is to disrupt debate. They tweet the same things, with some variation, so they are not detected by the Twitter anti-bot algorithm, and over time they multiply their presence and give the impression to the Twitter audience that it’s not just one person with this opinion, but loads of people. Which in turn is interpreted by the Twitter audience as significant or important.”

Researchers at the University of Oxford’s Comprop division have recently coined the term ‘computational propaganda’ to describe what is happening. What might have been thought of as sophisticated ‘micro targeting’ tools are being used maliciously in the war to influence public opinion online. At worst, they deliberately disseminate lies – and a botnet will respond by furiously retweeting and liking what is said.

Numerous anti-choice profiles on Twitter accurately fit the bot description. These accounts generally have a similar pattern: women’s names, with a generic ‘pro-life’ profile picture (eg. a mother or disability activist) and geo-tagged in counties including Louth, Galway, Mayo and Dublin. They have a high retweet ratio that would be difficult for human users to maintain without some level of account automation.

Fleming has encountered these on numerous occasions, and even before joining @repeal_sheild, began to keep track of their activity.

“They are trying to present themselves as women to gain credibility in the debate,” she notes. “They’ll use a term like disability activist, but if you check the time-line, there will be nothing about the UN charter for disability rights for example, or anything even remotely related to the cause. The only signifier in their timeline is sometimes an image of a young toddler with Down Syndrome that Down Syndrome Ireland have requested that people don’t use.”

PROPOGANDA WAR
Upon further inspection of one specific pro-life account with the handle @MarieAFlaherty, Hot Press discovered that the profile claims to be in Dublin but has a VPN address in North America, further intensifying worries among pro-choice activists of interference by American ideologues.

“There is no question about the link between American anti-choice organisations and Irish ones,” says Fleming. “They have very well-established links; they will post on their Facebook and Twitter accounts that they are flying over to Ireland. There are travelling for training and campaigning. We’ve heard a lot of people tell us that they’ve been canvassed by ‘pro-life’ Americans on the street.”
As we count down the days to the abortion referendum in May, the anti-choice side have been heavily focussed on carrying out a campaign of online warfare.

They have bought pro-repeal domain names such as repeal8th.ie, repeal8.ie and repealeight.ie, which upon clicking redirect you to an anti-choice site, thisismylife.ie. They’ve enlisted the help of an app company called u-Campaign, whose previous clients include gun lobbyists in America and anti-gay marriage activists in Australia, along with a political consultancy agency known as Kanto.
Ellen Coyne, senior journalist with The Times Ireland, has been keeping track of these groups in relation to the Eighth campaign.

“Kanto has been linked with some of the really negative aspects of the Brexit Leave campaign,” she says. “People probably saw that a lot of MPs were targeted via Facebook ads. It’s still very early days and we’ve yet to see if these tactics will emerge in Ireland, but I would safely anticipate that they will.”
As the tremors from fake news, the alt-right and the Brexit referendum are felt here, there have been moves by Fianna Fail TD James Lawless to criminalise the use of bots in political discussions online. The Online Advertising and Social Media (Transparency) Bill would set restrictions on online political advertisements and require purchasers to display a transparency notice.
However, the regulation of bots has been a controversial topic in many other countries, where concerns over the open internet and definitions of what constitutes a bot are debated. Segesten has heard these concerns many times before.

“It’s one thing to have these lofty ideals about transparency,” she says, “and maybe it’s a signal from the politicians’ side saying we are not going to tolerate a propaganda war like we have seen in the US. In reality, it is probably not going to be effective because of the hybridisation of accounts. How can you tell what is a bot or a cyborg? You cannot deny the existence of a human expressing their opinion and if they use technological advances to diffuse their opinion, is this illegal? No, of course not. So where is the line drawn?”

There may not be any easy answers, but this much is certain: where propaganda is concerned, the anti-choice gang will stop at nothing...