Wikipedia, one of the most used but sometimes unreliable resources on the internet, may be getting valuable help from Meta, which recommends the use of a new AI-based tool. – Sphere – to fix virtual encyclopedia entries. At present, mistakes and clearly malicious edits by well-meaning people are folklore.
A tool that uses a globe Database with over 134 million web articles to guide verification-dependent platforms against fake news. What makes this AI different is that it uses an inexhaustible store of information on the open web as a source, rather than traditional Google-like search engines.
How does Meta plan to fix Wikipedia?
In fact, Meta does not partner with Wikipedia in any way. The Sphere project is still in the research phase and is by no means implemented in the multilingual collaborative writing encyclopedia. But this is how the Menlo Park giant sees WikiProject: The perfect lab for your mega editing tool.
Volunteer human editors are often incompetent to run the site. Receives more than 17 thousand articles per monthMeta has developed its AI model that promises to automatically scan citations at large scale to verify their accuracy. The tool even suggests alternative citations when few sources are encountered.
To get an idea of what this represents, when reviewing Wikipedia citations, keep in mind that its editors do this purely on the basis of common sense and experience. But in the case of Sphere, AI uses a “natural language understanding” (NLU) transformation model that simulates the way humans think. aim find a single resource To validate each statement
Source: Tec Mundo
