Partisan Twitter bots distorting U.S. presidential candidates' popularity - Action News
Home WebMail Tuesday, November 26, 2024, 01:05 PM | Calgary | -8.3°C | Regions Advertise Login | Our platform is in maintenance mode. Some URLs may not be available. |
WorldVideo

Partisan Twitter bots distorting U.S. presidential candidates' popularity

It may sound a bit childish, but bot (automated) Twitter accounts like @loserDonldTrump and thousands of others are changing the conversations around this election. They might even be changing minds.

Thousands of automated accounts known as bots flood site with messages for and against candidates

Thousands of Twitter bots (automated accounts) are flooding the site with messages both for and against Donald Trump and Hillary Clinton, seen here in statuette form in a Naples, Italy, shop on Tuesday, Oct. 18, 2016. (Cesare Abbate/ANSA/Associated Press)

@loserDonldTrump wasn't born yesterday. It was actually the day before.

Just after 3 p.m. ET on October 18, the baby bot burst into the Twitterverse. Someone, somewhere, who is clearly not a fan of the Republican presidential nominee, wrote a piece of code for a Twitter account that would fill people's feeds with anti-Trump messages.

It all sounds a bit childish, but @loserDonldTrump and its thousands of cranky cousins are changing the conversations around this election. They might even be changing minds.

The brand new baby bot started with a few hundred posts an hour, and by the end of Wednesday night's presidential debate it had retweeted thousands of times, earning itself the dubious honour of the busiest bot of the night, according to social media analytics firm Cision Canada.

It far outpaced the next contender, American Right Now (@amrightnow), a conservative pro-Trump bot, or automated account, devoted to posting conspiracy theories, which on debate day created 1,200 posts. @loserDonldTrump more than doubled that output.

Cision's James Rubec talks proudly of "watching the birth" of this bot. Seeing its first retweet, he and his team set to work trying to figure out its story. Turns out, it wasn't complicated.

The bot takes all tweets that mention @realDonaldTrump (Donald Trump's verified Twitter handle) with the word "loser" or #loser.

"It finds those and shares them indiscriminately. It will share every single post," Rubec says.

Bots seem like real people

On this, the day after the debate, the bot seems a bit tired. Rubec says it's changing its pattern. It appears to be taking a break for an hour, then sending out a flurry of 120 posts the next hour. It's like doing high-intensity Twitter interval training. Only, Rubec suspects the clever bot is trying to mix it up in order to avoid being suspended by Twitter, which is trying to clean its ranks of bots.

James Rubec, of social media analytics firm Cision Canada, says the volume of tweets from bots (automated accounts) can make a candidate appear more or less popular than they are in reality. (CBC News)

When you think about these accounts in isolation, they can seem harmless, like distractions. But the combined noise of all the political bots might just be altering the tone of political talk.

Rubec explains their power as one of amplification. These automated accounts can seem real enough, with pictures of real people and posts that certainly sound like they are human, as well as geolocators that appear to place them in the U.S. Some are so convincing that people real people get into prolonged arguments with them. Frustrating enough fighting with humans online. But battling a bot on Twitter? No amount of wit will win that one.

Add the air of legitimacy to the sheer volume of automated posts, and these bots can make causes or candidates appear more or less popular than they are in reality.

"The bots create a false sense of the size of the conversation," Rubec says. "They will make certain things seem more appropriate to share on Twitter or on any social platform or online community."

The @loserDonldTrump Twitter account was the busiest bot (automated account) on the social media site on debate night Wednesday, according to social media analytics firm Cision Canada. (Cision Canada)

What sort of volume is out there? It's pardon the phrase huge. And it's especially huge in favour of Donald Trump.

Philip Howard, a professor of internet studies at Oxford University's Internet Institute, estimates that fully a third of all the tweets in favour of Donald Trump and there are millions of them are generated by automated accounts. Both Howard and Rubec believe that some accounts can pump out as many as 10,000 pro-Trump posts a day.

Meanwhile, automated accounts appear to be responsible for only a quarter of pro-Clinton tweets. And in raw numbers, there are far fewer of those.

At its most basic, says Howard, automated posts in favour of Trump tend to outpace those in favour of Clinton at a rate of four to one.

"Trump has more bots sort of working to spread his message than Hillary has to spread her message out," says Howard, who, like James Rubec, spends days and nights staring deep into the data. "These [pro-Trump] bots," he says, "often tweet outrageous messages so falsehoods and lies and accusations. They tend to be a tool for negative campaigning."

To be clear, the bots may be functioning in favour of one candidate or the other, but it's not always easy to know who's driving them, or even paying someone to build them. All of this can happen without the candidates, or even the campaigns, knowing anything about it.

The automated Twitter traffic follows a curious pattern. Howard's sense is that the bots have shared more content in the days leading up to and the day after each of the three debates. During those mercurial 90-minuteface-offs, though, it sometimes seems like the bots back off a bit, let the humans do the talking and the screaming.

When they're tired out, the likes of the baby bot born barely 48 hours ago gear into action.

Bot Tweets Influencing U.S. Election

8 years ago
Duration 2:22
Automated accounts are sending out tweets that could amplify and distort the U.S. election noise.