A while ago, I came across an interesting article published in The Guardian, entitled: Could language be the key to detecting fake news? The proliferation of deliberately misleading speeches, articles, websites and social media posts has led to many researchers trying to work out what the linguistic characteristics of fake news are. The article mentions that several researchers found that fake news may have in common the language they use – fake news articles use more words related to sex, death and anxiety, and overly emotional language is often deployed. The researchers noted that propaganda tends to use abstract generalities like “truth” and “freedom.”
But how accurate can artificial intelligence (AI) be in telling fact from fake? One would argue that as AI becomes better at spotting fake news, so will bots in their editing of fake news in a way that helps to escape identification. A clever publisher of fake news will dress up the stories to suit the target audience.
Can the branding/reputation of a news source be an incentive to be honest or trustworthy? Perhaps. Apart from the fact that news companies are often government controlled, most are there to make a profit, which also raises another question: real reporting takes time and resources, while fake news can be produced almost free of charge.
Yes, we can evaluate one writing to see if it is more informal than we would expect. We can check if it contains lots of superlatives and emphatic language, however, still the best way to fight fake news is to be open-minded and diversify our sources of information. Regretfully, most people believe it, because they want to believe it. What matters now is not what is true, but what is believed to be true.