This is the scariest sentence you will ever read about the internet

For those of us who care about democracy and the importance of a healthy political discourse, a story published by Bloomberg this week, “How to Hack an Election,” was incredibly distressing.

In it, Colombian hacker Andrés Sepúlveda comes clean about eight years he spent allegedly using dark, and often illegal, computer skills to help conservative candidates throughout Latin America; some won and a few lost. He describes exploits in major elections in Colombia, Nicaragua, Honduras, Venezuela, Costa Rica, Panama and Mexico.

In his most noteworthy alleged victory, he claims to have assisted the increasingly unpopular Mexican President Enrique Peña Nieto win his 2012 election. Sepúlveda claims that he was given a budget of $600,000 to rig the election in favor of the Institutional Revolutionary Party (PRI) candidate. He says he installed malware in his opponents’ routers, which let him tap their phones and computers; sent prerecorded messages to tens of thousands of people in a critical swing state at 3 a.m. on election night, purporting to support another candidate to anger voters; and set up fake Facebook accounts of gay men who purported to support a conservative Catholic candidate, angering many. Bloomberg says that it has verified some of what Sepúlveda claims to have done, but there is no way of verifying it all.

Finally, what most caught my eye, and what I think has the most relevance in our broader society, Sepúlveda claims to have maintained an “army” of over 30,000 fake Twitter profiles to “shape discussion” around certain topics in favor of his candidate, “priming the social media pump with views that real users would mimic.”

This was, according to the report, the tactic that likely had the most impact. It’s a technique that academics call “cognitive hacking”—in which an attacker attempts to change people’s perception of reality. With the bots in place, Sepúlveda could fabricate trends, and “manipulate the public debate as easily as moving pieces on a chessboard,” Bloomberg writes.

Which leads me to the scariest quote I have read in a very, very long time, from Sepúlveda:

“When I realized that people believe what the Internet says more than reality, I discovered that I had the power to make people believe almost anything.”

What’s notable about this tactic is that it’s not even illegal, only against Twitter policy.

You may be wondering why Sepúlveda would admit to all this. Well, he’s telling the story from behind bars. Sepúlveda is currently serving 10 years in Colombian prison “for charges including use of malicious software, conspiracy to commit crime, violation of personal data, and espionage, related to hacking during Colombia’s 2014 presidential election,” writes Bloomberg.

Maintaining Twitter bots may not have been his major crime in the eyes of the law, but the impact that trend bots could have is enormous. Our generation has told ourselves that the value and beauty of social media is that it offers authentic, real-time perspectives from real people. It’s a way to get unfiltered news, rather than being spun by political pundits on cable television.

But that perception means we’re vulnerable to the dubious use of such technologies. And people in power have figured that out. In 2014, there was the “Cuban Twitter” scandal, in which USAID, an international development arm of the federal government, quietly created an app that allowed Cubans to message each other. Through the use of casual surveys in the app, its creators hoped to identify which users were pro-U.S., seemingly in hopes of giving them an outsized influence over their peers.

“When I realized that people believe what the Internet says more than reality, I discovered that I had the power to make people believe almost anything.” — Andrés Sepúlveda, Colombian hacker

The app ultimately failed, having got too big, too fast, and it became clear that the creators were unable to explain to Cuban authorities how it was paying for itself.

A more successful example of cognitive hacking was the Joseph Kony video that went insanely viral in 2012, which introduced the world to a Ugandan war criminal and inspired the U.S. Senate to write a resolution condemning him. The video, which was later found to have falsely manipulated the narrative, was eventually revealed not to have been a completely organic viral success but rather masterfully marketed through a series of seemingly unrelated religious groups that were actually connected and engaging in “stealth evangelism,” as one critic put it.

Speaking specifically about how Twitter armies affected the 2012 Mexican Presidential election, Andrés Monroy-Hernández, a researcher at Microsoft Research who studies the use of social media in Mexico, was dubious about their effects in an interview with my colleague Kashmir Hill in 2014.

“No one was fooled by the bots. People had a strong negative reaction toward them,” he said then. “But social media plays an outsized role in affecting other types of media — newspapers, radio, TV — and the sheer noise can fool some people into thinking a candidate is more popular than he is. Whenever something happens in the offline world, journalists head to social media to find out more.”

Going back to at least 2010, the use of Twitter bot armies have been decried for their ability to influence the outcome of elections in the U.S.. That year, an army of tweet bots persistently  attacked Massachusetts U.S. Senate candidate Martha Coakley, in what researchers at Wellesley College dubbed “Twitter bombs.”

The result was “disproportionate exposure to personal opinions, fabricated content, unverified events, lies and misrepresentations that would otherwise not find their way to the first page (of Google search results),” researchers wrote. Coakley lost the race.

In the latest Pew Research Center survey, 14% of all people polled said they primarily get news about the 2016 presidential election from social media, a tie with local television. Cable television leads the pack, with 24% of all surveyed. But among 18- to 29-year-olds, social media is the top source, with 35% of them saying it’s their primary news source. That makes young people much more susceptible to this kind of manipulation.

Not surprisingly, there’s a prime contemporary example for why we should be more questioning of the “data” and the “trends” that we are currently seeing ripple through the political sphere.

For news organizations, one of the biting facts of covering this election cycle has been that the words “Donald Trump” attract readers and watchers, thus generating advertising revenue, and incentivizing those same outlets to give him more exposure.

It has been mystifying, yet patently real. No one seems to like the guy; his unfavorability rating is currently at 63.3%, according to the Huffington Post. Yet traffic to pieces about him and searches for his name have been hitting the ceiling since very early on.

This quote from Quartz, from January, explains it:

“We first noticed how Trump stories drove traffic—and thus revenue—at TheStreet.com last summer, when the name “Trump” suddenly spiked in Google searches. Traffic to stories about Trump, even if they contained no substantive new information or a comment from the candidate himself, was running some 10 times higher than that of any other candidate.

“Trump interest has since continued to dominate the web. A simple glance at Google Trends shows Trump searches far outpace those for any other candidate, Democrat or Republican (a phenomenon clearly at play during the Jan 28. debate).”

When media get a lot of traffic to a certain kind of story, it’s a sign that their audience is interested and they tend to write more about it. So, if you wanted to manipulate journalists into writing about something, you could send fake traffic to stories on that topic, in hopes that they write more.

Now, that’s not to say Trump is rigging the computers. There’s such a thing as a hate search, and a hate read, and I don’t doubt for a second that many people who hate Trump are reading about him just to keep up-to-date and informed on their hatred. But can that account for all the hype?

In yet another troubling quote from the Bloomberg piece, the authors ask Sepúlveda a blunt question: do you think the U.S. election is being tampered with? His response leaves no room for doubt.

“I’m 100 percent sure it is,” he said.

Daniel Rivero is a producer/reporter for Fusion who focuses on police and justice issues. He also skateboards, does a bunch of arts related things on his off time, and likes Cuban coffee.