Fake news is a ‘Big Data’ problem of massive proportions. It is an issue that is of great concern to many organizations and governments worldwide. Fake news websites deliberately publish hoax news, propaganda information and disinformation— using social media networks to disseminate the information. There are so many reasons why fake news exists today, but this article will not focus on the cause, but rather the cure.
Fake news is an internet disease of some sorts, and is a data veracity problem. It doesn’t matter the amount of data an organization processes if the data is incorrect. Fake news has tarnished the image of several social media sites as valid news sources and will continue to do more harm if not checked.
In France, the coming presidential election has been awash with fake news stories on social media. Researchers from Oxford University discovered that a quarter of the links shared on Twitter were from fake news stories. These fake political links were identified as a deliberate attempt to spread hyper-partisan and extreme views to the French public.
How can we fight fake news?
This is a question that is asked by many governments and organizations today. Manually sifting through data to check its authenticity is impossible so naturally Artificial Intelligence (AI) is the best way to fight Fake News.
Organizations can use AI technology to process natural language, identify semantics, and for machine learning processes. These techniques can provide a clue about how authentic a data set is.
Here’s an example. Let’s say news erupts on social media that a two-headed monster was seen wandering around Times Square, a natural language processing (or NLP) bot can look at the headline of this fake news: the subject of the news, the geo-location and main content. The AI will attempt to discover if this news has also been reported on other social media websites, and the AI will weigh the facts against established media sources.
This method may not be 100% productive all the time, but it is better to lose some valuable information, than to allow fake news infiltrate communities and disrupt governance.
Facebook was accused of promoting fake news during the USA elections in 2016, and they have taken the fight to a whole new level. The smart folks at Facebook decided to use algorithmic machine learning to detect fake news and mix it up with user behavioral patterns. They designed a system to reduce click-bait by demoting stories that featured click-bait stories. In January 2015, Facebook launched a system that demoted heavily flagged links. This type of machine learning is a great start to build a bigger, bad ass artificial intelligence monster that will hopefully put an end to fake news on Facebook.
Google was also accused of promoting fake news during the US presidential election — and Google has stated that they are committed to fighting fake news. Earlier this year, they launched a fact checking tool, which will help in the fight against fake news. This tool is a big data analysis system. When users see fake news and click the fact check tool, they create a data set for easy analysis and classification.
Artificial intelligence and big data analytics is the way forward for many social media organizations, governments and large corporation who suffer loss of integrity and revenue from fake news. Although the process will be a tough one as AI’s may also classify real news as fake, I believed that machine learning systems will help make authentication more reliable over time.
Using pattern recognition is also another way that AI can be used effectively in this fight. An AI can detect patterns or words that are indicative of a fake news story. For example, a 19-year-old Stanford student known as Karan Singhal recently created a web extension called the Fake News Detector AI. This tool uses 55 different metrics to track factors regularly used in determining whether a news site is real or fake— like content and layout. This tool like so many other AI tools will need modifications and upgrades to eventually get it right, but no matter how long it take AI will win this fight.