A team of researchers from the universities of Jaén and Alicante, in Spain, have developed an application capable of automatically analyzing the news, determining its veracity with a high degree of precision .
Although the model is still in the testing phase, it is proposed as a really useful tool to help journalists and private readers to filter the large amount of information that is published and reaches us on a daily basis.
In order to identify fake news, scientists have developed false news and misinformation credibility inference models based on machine learning .
These artificial intelligence techniques offer the system the possibility of analyzing the news on two levels, to detect if the content structure coincides with that which any publication with journalistic rigor should have, and if there are inconsistencies.
The work has been published in the specialized magazine ‘Expert Systems with Applications’, which has been entitled ‘Exploitation of the discourse structure of traditional digital media to improve the automatic detection of fake news’.
In it, they present the prototype of a “fake news” detector for websites , whose main objective is to give readers greater confidence, and to provide journalists with new tools that allow them to distinguish between different pieces of information.
How does this system work?
The system works by analyzing the structure of the published news, especially taking into account traditional journalistic standards: the 5W + H rule and the inverted pyramid.
These references are based on the fact that a rigorous news story must contain information that answers the six basic questions (what, when, where, who, why and how); In addition, it must be presented in descending order of priority, from the most important to the finest details.
As the researchers state, the structure of a publication provides clues as to whether or not it has a journalistic basis , or whether, on the contrary, it is pretending to be a real news story.
From the analysis of natural language, scientists have developed an algorithm capable of detecting information that originally does not coincide with this structure, through calculations that are based on techniques of what we know as machine learning, so that the system “ learn ”as you accumulate more data.
In addition, the machine is capable of processing thousands of data simultaneously in a matter of seconds , something that a person could not carry out, so that journalists could with this system detect incorrect structures, compare and contrast sources, identify viral content or inconsistencies between the headline and the body of the news not only automatically, but immediately.
But it is a system that could also provide interesting benefits for the final reader, since the user will have clues as to whether or not the news they are reading meets certain standards.
The researchers carried out different tests on a data set, in Spanish, with more than 200 articles focused on health issues, of particular relevance today due to the numerous false news that have circulated and continue to circulate today about COVID -19.
Therefore, the objective pursued by the experts is to develop an application capable of automatically marking a certain text as ‘fake news’ while it is being read , and which alerts the journalist (or the reader) by means of a signal about which parts of The information may be false, indicating as a reference other similar useful texts to verify its veracity .
