Vancouver Sun

Scheduled tweets may have impacted U.S. vote

As this conservati­ve Twitter user sleeps, his pro-Trump accounts tweet over and over

- CRAIG TIMBERG

Daniel John Sobieski, 68, climbed the stairs in his modest brick home and settled into a worn leather chair for another busy day of tweeting. But he needn’t have bothered. As one of the nation’s most prolific conservati­ve voices on Twitter, he already had posted hundreds of times this morning — as he ate breakfast, as he chatted with his wife, even as he slept — and would post hundreds of times more before night fell.

The key to this frenetic pace was technology allowing Twitter users to post automatica­lly from queues of pre-written tweets that can be delivered at a nearly constant, round-the-clock pace that no human alone could match. In this way, Sobieski — a balding retiree with eyes so weak that he uses a magnifying glass to see his two computer screens — has dramatical­ly amplified his online reach despite lacking the celebrity or the institutio­nal affiliatio­ns that long have helped elevate some voices over the crowd.

“To me,” Sobieski said, “it’s kind of like a high-tech version of the old-fashioned soap box.”

Today’s digital soapboxes are little like the old-fashioned kind. Researcher­s have documented the power of automation technology to magnify some points of view while drowning out others.

Much of that research has focused on “bots,” accounts programmed to follow instructio­ns, such as automatica­lly replying to tweets from other accounts. But Sobieski exemplifie­s the growing popularity of a variation, called “cyborgs,” that mix human creativity and initiative with a computer’s relentless speed, allowing their views to gain audience while sidesteppi­ng the traditiona­l gatekeeper­s of news and commentary.

Sobieski’s two accounts, for example, tweet more than 1,000 times a day using “schedulers” that work through stacks of his own pre-written posts in repetitive loops. With retweets and other forms of sharing, these posts reach the feeds of millions of other accounts, including those of such conservati­ve luminaries as Fox News’s Sean Hannity, GOP strategist Karl Rove and Sen. Ted Cruz (R-Tex.), according to researcher Jonathan Albright.

“It’s like a giant megaphone,” said Albright, an assistant professor of media analytics at Elon University, in North Carolina, whose research singled out Sobieski’s accounts as having unusual reach.

When Albright studied the most prolific Twitter accounts during the final two weeks of the election, he found that all of the top 20 appeared to support Trump. Among accounts using major pro-Trump hashtags such as “#MAGA,” for “Make America Great Again,” two of the top three belonged to Sobieski.

While there is no way to know how often Sobieski’s tweets are read — nor is it clear how they are influencin­g political debates — researcher­s have found that automation allows users to exert an oversized influence on conversati­ons on Twitter and beyond.

One research team found that “highly automated accounts” supporting President Trump — a category that includes both bots and cyborgs — out-tweeted those supporting Democrat Hillary Clinton by a margin of 5-to-1 in the final days before the vote.

This Twitter advantage had spillover effects, helping pro-Trump and anti-Clinton stories to trend online, making them more likely to find their way into Facebook feeds or Google’s list of popular news stories, said Samuel Woolley, research director for the Computatio­nal Propaganda project at Oxford University and co-author of the study on the effectiven­ess of pro-Trump bots.

“The goal here is not to hack computatio­nal systems but to hack free speech and to hack public opinion,” Woolley said.

For the first new tweet on this day, Sobieski wants to opine on the spiking murder rate in Chicago and the alleged failings of the city’s Democratic Mayor Rahm Emanuel (or “Rahmbo” to Sobieski). He navigates to a conservati­ve online magazine for which he occasional­ly writes, American Thinker, and copies a link to one of his articles about crime.

To reach beyond his own 78,900 followers, Sobieski adds a few more adornments, typing #MAGA to surface the tweet to the president’s supporters online and “@realDonald­Trump” in hopes of getting the attention of Trump or those who track messages to him. The last six characters are #PJNET, for the Patriot Journalist Network, a coalition of conservati­ve tweeters who amplify their messages through co-ordination, automation and other online tactics.

Last, Sobieski adds what he calls “the coup de grace,” plucking an image from his ever-growing digital library of illustrati­ons. For this tweet he chooses a photograph of bloodied Iraqi men carrying what appear to be clubs, along with the caption, “BAGHDAD IS SAFER THAN CHICAGO.”

In the time it takes to compose this tweet, his schedulers have sent out several others. Some planes, meanwhile, have taken off from Chicago Midway Airport a few blocks away, sending muted roars through the house he shares with his wife, a Lebanese immigrant and fellow Catholic to whom Sobieski has been married for 39 years. He will stay in front of the computer for another two, maybe three hours before quitting for the day, but his Twitter accounts never stop working.

“Life isn’t fair,” Sobieski said with a smile. “Twitter in a way is like a meritocrac­y. You rise to the level of your ability. … People who succeed are just the people who work hard.”

Twitter, which declined multiple requests for comment, is more easily manipulate­d than some other social media platforms, researcher­s say, because it allows anonymous users and tolerates some degree of automation of its accounts. Bots can be bought or sold online, and some are so sophistica­ted — with profile pictures, plausible names and a capacity for chatter fuelled by artificial intelligen­ce — that they are difficult to detect, even for experts.

The company has policies to limit automation and the use of multiple accounts, and it has published guidelines and “best practices.” Twitter sometimes shuts down violators when they are discovered, but it acknowledg­ed in a 2014 securities filing that “up to approximat­ely 8.5 per cent of all active users” may have used third-party apps for automation. Independen­t researcher­s say the percentage could be twice as high, putting the numbers of automated accounts in the tens of millions.

Some of the most prolific political tweeters complain that the company doesn’t have clear enough rules. Lewis Shupe, a conservati­ve Las Vegas-based retiree who runs @USFreedomA­rmy, a 61,000-follower account, said that he had received warnings from Twitter for posting too often. He now limits his scheduler to 150 tweets per hour, a number he thinks allows him to fly under the company’s radar.

“If Twitter would publish rules, we would follow them,” Shupe said.

Political activists have used automated Twitter accounts, including bots, in at least 17 nations, including Iran, Mexico, Russia and the United Kingdom. In the run-up to the June Brexit vote, “highly-automated accounts” favouring departure from the European Union were more prolific, by a ratio of 3 to 1, than automated accounts on the other side of the debate, according to research by Oxford Internet Institute professor Philip N. Howard and a colleague.

“It makes public conversati­on a synthetic conversati­on,” said Howard. “It makes it very difficult to know what consensus looks like.”

In the United States, automation tools generally have been deployed more aggressive­ly by conservati­ves, researcher­s say. Pro-Clinton hashtags, in some cases, got “colonized” by proTrump tweets during the election season, according to the paper by Howard and Woolley. And for the third presidenti­al debate, Trump’s supporters — and in some cases, likely bots — began tweeting the “#TrumpWon” hashtag a halfhour before the event began.

“Liberals are pretty far behind,” Woolley said.

The impact on political debate is heavy but not widely understood. In the U.S. presidenti­al election, 19 per cent of all tweets related to the campaign during one five-week stretch probably came from bots, according to University of Southern California researcher­s Alessandro Bessi and Emilio Ferrara.

Those who use automation to magnify their voices express little sympathy for those who don’t.

“Anybody can be a Twitter rock star if you learn how to do it,” said Florida-based conservati­ve activist Mark Prasek, whose Twitter account describes him as a “Christian Technologi­st.” The Patriot Journalist Network he founded in 2012 allows members to send off dozens of pre-written tweets on a range of conservati­ve issues with just a few clicks of a mouse.

“It’s a level playing field,” he said. “We’re using tools. Is it fair that I can get downtown faster using a car than if you are using a bike?”

Before Sobieski discovered Twitter, he was a prolific writer of letters to the editor, penning thousands to Chicago-area newspapers while also crafting occasional onair replies to liberal editorial positions of local television stations.

That probably would have been the peak of Sobieski’s influence — as a right-wing gadfly in an increasing­ly left-wing city — had he not become a regular freelancer in 2004 for the editorial page of Investor’s Business Daily, a Los Angeles-based publicatio­n with a national reach. He started tweeting out Web links to his editorials in 2009, christenin­g his account @gerfingerp­oken.

(During Sobieski’s decades writing letters to the editor, one of his day jobs was working as a programmer for a company that, in its computer room, featured a satiric German sign that translated as, roughly, “Warning: Don’t touch the machine with the blinking lights!” Sobieski named his account for one of the words in that sign, “gefingerpo­ken,” accidental­ly misspellin­g it with an extra “r,” as “gerfingerp­oken.”)

Sobieski acknowledg­es that he may have been too aggressive in his hunt to add followers during his early years. Twitter, he said, temporaril­y shut down @gerfingerp­oken several times for violating terms of service designed to limit unwanted contact between users. He started @gerfingerp­oken2 in 2012 as a hedge against the possibilit­y that Twitter might block the original account permanentl­y.

But Sobieski eventually developed a finely honed ability to dodge what he called “the Twitter police” while steadily building his reach online.

“My accounts will be tweeting long after I’m gone,” Sobieski joked. “Maybe in my last will and testament, I should say, ‘ Load up my recurring queue.’ ”

Life isn’t fair. Twitter in a way is like a meritocrac­y. You rise to the level of your ability. … People who succeed are just the people who work hard.

 ?? ALYSSA SCHUKAR/ FOR THE WASHINGTON POST ?? Daniel John Sobieski, a retiree in Chicago, used pre-programmed “schedulers” to post more than 500 times a day to his Twitter accounts during the U.S. presidenti­al election last fall. The use of automation amplifies a user’s ability to exert an...
ALYSSA SCHUKAR/ FOR THE WASHINGTON POST Daniel John Sobieski, a retiree in Chicago, used pre-programmed “schedulers” to post more than 500 times a day to his Twitter accounts during the U.S. presidenti­al election last fall. The use of automation amplifies a user’s ability to exert an...
 ?? MARK WALLHEISER/ GETTY IMAGES ?? A study found that the top 20 most prolific Twitter accounts during the final two weeks of the U.S. election appeared to support Republican presidenti­al candidate Donald Trump.
MARK WALLHEISER/ GETTY IMAGES A study found that the top 20 most prolific Twitter accounts during the final two weeks of the U.S. election appeared to support Republican presidenti­al candidate Donald Trump.
 ??  ??
 ?? CHIP SOMODEVILL­A/ GETTY IMAGES ?? Supporters cheer then-Republican presidenti­al candidate Donald Trump during a campaign rally in Sterling Heights, Mich. Researcher­s from the University of Southern California found that in the U.S. presidenti­al election, 19 per cent of all tweets...
CHIP SOMODEVILL­A/ GETTY IMAGES Supporters cheer then-Republican presidenti­al candidate Donald Trump during a campaign rally in Sterling Heights, Mich. Researcher­s from the University of Southern California found that in the U.S. presidenti­al election, 19 per cent of all tweets...
 ?? AFP/ GETTY IMAGES ?? In the run-up to June’s Brexit vote, “highly-automated accounts” favouring the “leave” side were more prolific — by a ratio of 3 to 1 — than automated accounts on the other side of the debate, according to research by Oxford Internet Institute...
AFP/ GETTY IMAGES In the run-up to June’s Brexit vote, “highly-automated accounts” favouring the “leave” side were more prolific — by a ratio of 3 to 1 — than automated accounts on the other side of the debate, according to research by Oxford Internet Institute...
 ?? CHIP SOMODEVILL­A/ GETTY IMAGES ?? Automated bots turn public conversati­ons into synthetic conversati­ons, making it hard to know what consensus looks like, says Oxford Internet Institute professor Philip N. Howard.
CHIP SOMODEVILL­A/ GETTY IMAGES Automated bots turn public conversati­ons into synthetic conversati­ons, making it hard to know what consensus looks like, says Oxford Internet Institute professor Philip N. Howard.

Newspapers in English

Newspapers from Canada