Recognizing Fake News with Algorithms

Fake news has been a major topic for the last few years. According to Google Trends, the term made a giant leap in Google's worldwide search in October 2016 and has been a constant theme ever since. The term has even become part of everyday language. But what exactly is Fake News, why has it gained relevance right now and what can be done about it?

(Source: https://trends.google.de/trends/explore?date=all&q=fake%20news)

Fake news is the distribution of false or questionable information that is either completely invented or sold as factually correct news. Fake News was and still is associated with the 2016 US election campaign, and to this day the influence of Fake News had in this context has not been fully explained. Russian agents are said to have deliberately circulated false information in order to influence the outcome of the elections in Trump's favor. According to reports in the German weekly magazine "Zeit", the twenty most successful Fake News were shared, liked, and commented on more often than the twenty most successful articles by reputable news outlets.

The most effective distribution channels for Fake News over the last few years have been various social media platforms. Social media have an immense influence on the circulation of Fake News, since they have a large reach and because they can target individual users directly. Facebook, like the other social media, uses an algorithm to present the user with the content he or she might like best, with the aim of keeping the user on the platform for as long as possible. Hotly discussed articles or contributions are more likely to keep users engaged with the platform.

To be able to detect Fake News effectively and automatically, new algorithms are necessary. A data management company from Munich, for example, has developed a program that does not look at the content of the distributed news, but rather at the distribution channel through which the news is distributed. This involves feeding known False News into a machine learning algorithm and analyzing their propagation. In addition, accounts are identified that are often associated with the distribution of Fake News. In doing so, they established a model to assess whether a given article or contribution is true or whether it could be false information.

The German Research Center for Artificial Intelligence (DFKI) uses a different method: It analyses whether images potentially associated with Fake News have already been published on the Internet before. This way it can be assessed when the image first appeared and whether it was modified with image processing software. Subsequently, a text analysis takes place, which places the picture in connection with the text in a given post. If the context of the original image does not match the content of the text, the program raises an alarm and reports that it is most likely Fake News. This program is still under development, but it will be available as a browser addon in the future to help users detect Fake News.

Since you cannot only count on the government or private companies to shield you from Fake News, here are some tips to help you recognize misleading information:

• Question information critically, instead of just consuming it: Does the article sound credible? What are the sources on which it relies? In which medium was it published? Here are 10 key questions that can help you recognize Fake News.
• Use websites like org, politifact.org, factcheckeu.org or the YouTube Data Viewer to check information for its credibility.
• Download extensions or addons for you browser, which help you detect Fake News: S. Detector, Fake News Detector, Trusted News.