Meta is said to have developed an artificial intelligence (AI) model that can automatically scan Wikipedia citations on a large scale to verify their accuracy.
According to Engadget , the new tool can also suggest alternative citations when it finds a piece of poorly sourced text.
To evaluate citations, Wikipedia’s editors will rely on common sense and experience, and Meta’s AI will do a similar job, where it uses the “Understanding Natural Language” transformation model. nature” (NLU) to try to understand the different relationships of words and phrases in a sentence.
Meta’s Sphere database includes more than 134 million web pages that serve as knowledge indexes for the AI system. When it does the job of checking citations in an article, the model finds a single source to verify every claim.
Meta hopes the new model can help address misinformation issues on Facebook, improve content vetting efforts and increase the overall trustworthiness of online information. While the new tool improves, Meta hopes to build a platform that Wikipedia editors can use to systematically verify and correct annotations.
largest online encyclopedia the world’s . Recently, the Wikimedia Foundation has partnered with Meta to find ways to improve verification, especially as the site grows with more than 17,000 articles monthly, many of which contain incomplete, missing citations. or simply incorrect.