Fighting Fake News with Big Data
Big Data, Bots and Fake News: These terms of the digital world often emerged in the context of the US election outcome and the upcoming election of the Bundestag. What is the connection between those and can Big Data approaches help in the battle against fake news?
What are fake news?
You may have heard the term “Fake News” in connection with the US elections. Fake news are defined as false and misinformation, which are often spread via electronic channels and, especially, social media. They are created and spread by individuals or groups acting on their own or outside behalf. There are personal, political and economic motives for creating such false messages. Algorithms of various types and social bots play a central role in the dissemination, as well as user posts, likes and retweets.
A problem of the informational age?
At first, fake news were mainly related to Facebook and Twitter and other social media, where the content is published in text, pictures or via links. The term has become a political discussion and the underlying phenomena became a problem of our information society, in which manipulation and disinformation are increasing. Next year, Germany will elect a new Bundestag on September 24, 2017 – and the topic “Fake News” will certainly play an important role again. EU Commissioner Juncker also interfered in the debate about false Internet notifications. He appealed to corporate groups like Facebook to do more against Fake News. Credibility is “their most important asset,” he told the radio media group.
Companies like Google and Facebook had so far repeatedly withdrew their statement that they were pure infrastructure providers. The corporations are only slowly recognizing their own social responsibility and partially only after massive public pressure. Mark Zuckerberg is one of the most popular examples. He first thought that it is a “rather crazy idea” that Fake News are a real problem and then, a few days later, he listed specific possible countermeasures. The media and parties are creating special facilities to identify and eliminate fake news. Facebook, for example, has now engaged the foundation-financed research agency Correctiv to identify and correct falsehoods and lies in the Facebook network.
Solutions for fighting fake news
Authorities are expecting hacker attacks during the Bundestag election campaign and especially on the election day. Possibly also with the goal to infiltrate networks with Fake News. Technical measures had already been taken, assured the responsible person for the election, called Sarreither. He will inform quickly about fake news. For this purpose the infrastructure of the data center was expanded and they could switch computers and the location. In cases of an emergency, he would also use the cyber defense center of the Federal Government. He said that the Bundestag election is technically secured “so that it is protected against all manipulation attempts”.
Machines like Social Bots are often used for the spreading of Fake News, but they are also important for the defense of Fake News. Fake News can be supported by algorithms that recognize Fake News as such and warn people about them. These pattern recognition algorithms need a well-defined problem, but fake news are not clearly defined: Sometimes fake news are rumors, sometimes deliberate lies, sometimes misunderstood satire or irony. An algorithm that exposes conscious lies relatively well fails, when someone is believing these lies and reproduces them in his own words. In this case the algorithm does not use those predictors that the machine learning methods discovered in conscious untruths. If he believes that it is the truth, he is telling it with full conviction – even if it is not. Thus, only the context helps to recognize content clearly as wrong or correct. Nevertheless, the machines are statistically better than a human-being.
What is the best way to prevent the spread of Fake News? In summary a combination of humans and machines is probably the best remedy against fake news. Algorithms can help to alleviate the problem by exposing potential false messages and then asking people for advice on which of the messages are actually problematic. However, the critical thinking of the readers remains the central challenge.