Halfthe parent company that owns Facebook and Instagram has aartificial intelligence able to verify the accuracy of quotes on Wikipedia The automated software is designed to assist editors in their task of keeping the encyclopedia free of falsehoods and inaccurate information.
AI is called Side, and can translate words into numerical values. In this way, the model can identify any contradictions between different paragraphs and assign a degree of probability to each quote. In the event that there is something suspicious, the artificial intelligence reports the problem to the editor, who can investigate and decide whether it is necessary to intervene and correct the voice.
No small help, as about 17,000 new ones are made every month entries on Wikipedia For real editors, almost all volunteers, it is almost impossible to keep a constant eye on every page. Automation can therefore play an extremely valuable role in the fight against disinformation. Meta shared a video illustrating the capabilities of its AI Side.
In a press release, Meta explains that it plans to deploy its AI with the aim of strengthening the authority and quality of the pages on Wikipedia, a tool “practically everyone uses”. Moreover, according to Meta, Side could help the research community “with the most difficult challenges related to the development of artificial intelligence”. The company added that the model was created using one of the most complete and comprehensive datasets ever used to design a Natural-language Understanding (NLU) system.
Source: Lega Nerd
