Democracy Fails the Turing Test

The spread of fake news in the 2016 election shows that Americans need to be more sophisticated about what they read.

April 28, 2017

The views expressed are the author's own and not necessarily those of the Brennan Center for Justice.

Alan Turing is the British mathematician Benedict Cumberbatch portrays in the film, “The Imitation Game.”  Turing is credited with creating one of the first programmable computers, which broke the Nazi’s cryptography during World War II. Turing’s work was critical for the Allied victory. He is widely considered the father of artificial intelligence and theoretical computing. And, in an impossibly cruel twist, Turing pleaded guilty to indecency charges for a homosexual affair he had after the war. He received hormone injections for a year, permanently lost his security clearance, and was denied entry to the U.S.

He is also the creator of the “Turing Test” which is a method for determining whether a machine’s communications are indistinguishable from that of a human. If the human interacting with the machine can’t tell whether they are communicating with a machine or another human, then the machine has passed the Turing Test. It wasn’t until 2014 that a computer passed the Turing Test for the first time.  It’s worth noting that “the computer” in this case was a Russian-designed chatbot named Eugene that had the persona of a 13-year-old boy.  

The old joke about the internet used to be, “on the internet, no one knows you are  really a dog.” Now the joke should be, “on the internet, no one knows you are really a chatbot.”

Chatbots had an interesting moment when Microsoft released its chatbot named “Tay” in March 2016.  Tay was an experiment in “conversational understanding.” The idea was that the more a user chatted with Tay, the smarter it got, resulting in “casual and playful conversation.” That’s one way of looking at it. Tay lasted all of 16 hours. The chatbot was overwhelmed by users manipulating Tay to spout “playful” expressions sexism, racism, and anti-Semitism.

Remember that March 2016 was in the midst of an election year. When users asked Tay to “repeat after me” Tay parroted back claims from Donald Trump such as “WE’RE GOING TO BUILD A WALL AND MEXICO IS GOING TO PAY FOR IT.”

As the House and Senate Intelligence Committees investigate the role of Russian influence in the 2016 election, they are likely to look at allegations that “thousands of Russian bots…pump massive amounts of disinformation and harassment into public discourse.”

One of the few outlets to cover the threat of Russian hackers was Samantha Bee’s satirical news show “Full Frontal.”  Bee traveled to Moscow and interviewed Russian hackers wearing ski masks who matter-of-factly admitted they were trying to manipulate the American election, including trying to reach “simple people” who are “lazy and believe everything they read…” As we now know, one of the approaches of the Russians was to discredit the legitimacy of the election.  

The noise created by Russian hackers and chatbots was far-reaching. A Stanford study released last month concluded that the average American saw at least one fake news story during the election and that many saw far more. (Perhaps even more troubling is that of those respondents who could recall a specific fake news story, about half of them believed it.) As Former FBI agent Clint Watts testified in before Senate Intelligence Committee, “[i]n the lead up to the 2016 election, fake news stories were consumed at higher rates than true stories.” 

Of course, not all fake news was generated by Russians operatives. There was plenty of fake news generated by Americans engaging in good ‘ol fashioned capitalism by posting click bait.

And whether he realized it or not (at this point I’ll give him the benefit of the doubt without more evidence of any venality), candidate Trump disseminated electronic propaganda from all sources via his Twitter feed to his millions of followers.  Perhaps the most notorious of them is Trump’s utterly fallacious November 27, 2016 Tweet: “In addition to winning the Electoral College in a landslide, I won the popular vote if you deduct the millions of people who voted illegally.”  

More disturbingly, Watts testified that the Russians are still tweeting at President Trump, hoping to catch his attention. Of course, Trump did not act alone. Fake news from Russia and elsewhere spread like a contagion as it was shared and re-shared.

As the FBI continues with its parallel investigation into Russian interference in the 2016 election, we need to ask why these measures were so effective?  For example, when a bot pretending to be a “Bernie Bro” wouldn’t budge on any point, why did some think it was a real Bernie Sanders supporter and not a bot that had been programmed to mess with us? If American democracy is going to weather the next tidal wave of fake news, we need to get better at the Turing test.