A group of Brazilian researchers has created a web platform that can automatically detect incorrect information on the Internet.
Developed by scientists at the Center for Mathematical Sciences Applied to Industry (CeMEAI), the system uses a combination of statistical models and machine learning techniques to determine whether a given piece of content is likely to be incorrect in Brazilian Portuguese. Initial tests indicate that the platform is able to detect fake news with an accuracy of 96%.
CeMEAI is a research center located in the Faculty of Mathematics and Informatics of the University of São Paulo in the city of São Carlos in the state of São Paulo. The center is supported by grants from the São Paulo Research Agency (FAPESP).
In an interview with the FAPESP news agency, the project coordinator and head of technology transfer, Francisco Louzada Neto, said the aim of the project is “to offer society an additional tool to not only subjectively recognize whether a message is false or not”
;. .The system uses statistical methods to analyze writing characteristics such as words used or more commonly used grammar classes. These are then fed into a machine learning-based classifier capable of distinguishing language patterns, vocabulary and semantics of fake and real messages and automatically inferring whether the content submitted to the platform is fake.
The models were trained on a huge database of real and fake news and exposed to the vocabulary used in over 100,000 articles published in the past five years. The researchers will aim to use the fake news related to the upcoming presidential elections as well as content related to the Covid-19 pandemic to further calibrate the models.
The researchers also commented on the potential risks of the system in the interview, including the possibility that the system could be used by fake news creators to assess the potential for false content to be deemed genuine before it is published. “It’s a risk we have to deal with,” Lozada noted.
Earlier this month, Brazil’s Supreme Electoral Court announced it had signed agreements with eight of the country’s largest social media platforms to thwart disinformation and the spread of fake news that could adversely affect October’s presidential election.
Twitter, TikTok, Facebook, WhatsApp, Google, Instagram, YouTube and Kwai have all signed individual agreements with each platform, outlining the actions each platform will take to stop the spread of false and misleading information. However, LinkedIn is still negotiating terms with TSE, the country’s highest electoral justice body. Telegram remains unreachable despite attempts to contact the TSE.