A machine learning system aims to detect fake news based on the understanding that the most consistent methods to detect both fake news and biased reporting are to look at the common linguistic features across the source's stories, including emotion, complication and construction
Researchers from MIT's Computer Science and Artificial Intelligence Lab (CSAIL) and the Qatar Computing Research Institute (QCRI) developed this system in collaboration with the Hamad Bin Khalifa University in Qatar. They took data from a website with human fact-checkers who analyze the accuracy and biases of more than 2,000 news sites.
They then fed that data to a machine learning algorithm called a Support Vector Machine (SVM) classifier, and programmed it to classify news sites the same way. When given a new news outlet, the system was then 65 percent accurate at detecting the level of "factuality," and approximately 70 percent accurate at detecting the political leaning.
The researchers also produced a new open-source data-set of more than 1,000 news sources, explained with factuality and bias scores, the world's largest database of its kind.