“Falsehood flies, and the truth comes limping after it,” wrote Jonathan Swift in 1710. Variations of the saying, including the punchier "A lie gets halfway around the world before the truth has a chance to get its pants on,” have been (likely incorrectly) attributed to everyone from Winston Churchill to Mark Twain, but the timeless truth remains: it’s amazing how catchy an outright falsehood can be.
Since at least the days of the newspaper wars between Joseph Pulitzer and William Randolph Hearst, on some level less-than-scrupulous media outlets have understood that our faster, more reactive brain systems bias us towards internalizing more emotional, good-vs-evil stories—especially when the “good” and “evil” in question align with our own prejudices.
For instance, in the late 1890’s, Hearst’s papers whipped up a frenzy of anti-Spanish sentiment, characterized by unproven claims, propaganda, and bold-faced lies. “When the USS Maine exploded and sank in Havana Harbor on the evening of 15 February 1898,” writes The Public Domain Review, “huge headlines in the Journal blamed Spain with no evidence at all.” The Spanish-American War soon followed.
The internet may have democratized communication, but the growing lack of gatekeepers makes it harder than ever to sort the truth from the lies—and the lies seem to be everywhere. Log into Facebook, and check the feed of your politically extremist relative (we all have one), and you’re bound to see some modern examples of “yellow journalism” pop up on your screen.
“Now, the only thing that could make this story even more insulting or outrageous were if any part of it were actually true,” deadpans popular Youtuber Shaun in his video “Outrage News.” He then proceeds to, point by point, carefully debunk a series of sensationalist claims, using only facts and reason.
Watching the truth slowly, methodically unfold is deeply satisfying, but it is time-consuming to research and refute every single bold claim you come across. Technology—be it the printing press or the internet—got us into this mess. Surely it should have the tools to get us out of it, too.
Good news: researchers from MIT’s Computer Science and Artificial Intelligence Lab and the Qatar Computing Research Institute (CSAIL and QCRI, respectively) have recently teamed up to bring us a new system which uses machine learning to help determine the accuracy or bias of a news story. Instead of focusing on proving or disproving individual claims, the system attempts to judge the integrity of the news source itself. After all, a group comfortable printing one lie is more likely to print another.
From analyzing just 150 articles from any given source, the system can gage the relative reliability of a news outlet, with an accuracy of roughly 65%. The system can also guess in which political direction the bias leans about 70% of the time.
News hounds take note: the system’s main tools to determine a source’s trustworthiness are the types of language used. When articles repeatedly used hyperbolic phrases and loaded, emotional terms, the source was more likely to be questionable. Outlets with a left-leaning bent stressed language of harm/care and fairness/reciprocity, while the right-leaning ledgers stressed concepts like authority and sanctity.
So the next time you come across an article that feels too outrageous to be true, remember: maybe it is.