Why AI is a Valuable Technology for Detecting Fake News by Objectively Categorizing Online News Articles

Jeff Finn, CEO of zvelo, explains why AI is a valuable technology for combating the “fake news” epidemic through its ability to quickly and objectively categorize online news articles based on how likely they are to present information accurately.

*****The following article, by Jeff Finn, appears as a content marketing article on MarTech Advisor’s web site and was originally published on May 26, 2017.

If you’re a consumer of digitally disseminated content (and by reading this then you are precisely that), you may be finding it increasingly hard to believe what you’re reading. The issue of Internet users consuming, believing, and sharing false information has become so dire that analysis suggests the most popular fake news stories outperformed the top true stories in total engagement during the final month of the 2016 election. Further investigations have revealed that the rather unsettling proliferation of fake news sources isn’t even motivated by politics all that often, but instead by money. In interviews, the non-political creators of fake new sites have described the perverse incentives driving them to generate fictitious information, and the substantial monetary rewards they reap thanks to ad revenue.

Fake news is now a concern being acknowledged and confronted by major tech industry players like Google and Facebook, which play central roles in the phenomenon (Google recently announced it will ban fake news sites from its AdSense advertising service, while Facebook has updated its ad policy as it weighs new methods for user reporting of dubious news stories). The government of Germany is debating its options as well, working on legislation that would impose steep fines on social media firms that don’t delete fake news within 24 hours of being posted. Of course, these bans and fines are potential free speech violations if not carefully applied – meaning that transparency and objectivity must be at the center of any proposed solution.

Assuming no one likes to be lied to, and that the accuracy of the information we consume is crucial to both our personal decision-making and society as a whole, individual Internet users may seek methods for avoiding fake news. However, the reality is that analyzing and verifying the accuracy of an online article isn’t so simple, especially for us human beings with limited time and resources. This is why it makes the most sense to leverage technology for assistance in vetting information online. Artificial intelligence, capable of performing fast and objective review of articles and the sites delivering them, is the perfect candidate for the job of categorizing such articles by their veracity (or lack thereof).

By extracting the semantic content of an article’s text to a knowledge base, and by using machine learning to look for key indicators, AI can quickly estimate what categories to which an article belongs. AI can then assist humans by conveying if an article appears to be trustworthy, biased, or entirely false, or if it belongs in such categories as satire, conspiracy theory, junk science, hate, clickbait, and so on.

Specifically, here are steps that an AI designed for fake news detection could take in performing an analysis of article content: Read more