Could language be the key to detecting fake news?

Is this the era of Fake News? We could say that fake news are one of the most prominent (and dangerous) phenomena on social media. Since they are spreading to traditional media as well, having a reliable way of identifying fake news is more relevant than ever.

It’s not a new situation: we have had fake news all the time, since we were in school. Yet, what has now changed is the scale and the concentration in time, and that their goals are very ambitious for example, they might be trying to influence elections or policies, or discrediting countries as a form of cyberwarfare among states and source of public relations crisis.

The popularity of social media means that fake news can globally spread in an instant and it's often difficult to find the source of a story after it has been widely shared. Companies that do fact-checking such as Newtral, Snopes or BuzzFeed can only address a small portion of the most popular rumors.

Fake news are serious, disruptive and damaging. The issue is so relevant that Facebook has 15,000 people working on detecting them manually. Having a reliable way of identifying fake news is essential, and technology is the only solution to do it. The first step is identifying who benefits from them, who is criticized, etc. For this reason, linguistics plays a key role when it comes to identifying patterns in the language used (with event extraction, for example, at a very fine-grained level) to flag it as potentially dubious. The linguistic characteristics of a written piece can tell us a lot about the authors and their motives. Purveyors of disinformation can be caught out by the particular words they use, according to a new research.

Ultimately, we may have to rely on artificial intelligence to do the heavy lifting for us: AI should be able to tell us whether large datasets of fake news contain or not those telltale linguistic patterns, which are invisible to the “naked eye”.

It's a public challenge, and it was the subject of discussion of the Mega Meetup organized by LUCA and Telefonica, where our CEO, Antonio Valderrabanos, participated today in Madrid as an NLP expert and jury member of the Data Science Awards Spain 2019.  

Trasera Luca



For more information, visit our website and follow Bitext on Twitter or LinkedIn.


Subscribe Here!