Meta’s AI Can Scan Wikipedia

Facebook claims it wants to assist in resolving the issue of false information spreading like wildfire on the internet—a problem it may have contributed to in the first place.

On Monday, Sphere, a new AI-powered product from Facebook parent company Meta, was introduced. It aims to assist in identifying and addressing “false news” on the internet. It is “the first [AI] model] capable of automatically examining hundreds of thousands of citations at once to determine if they actually support the relevant assertions,” according to Meta.

The statement follows years of criticism regarding Facebook’s own role in fostering and hastening the spread of internet disinformation around the world. The research team at Meta estimates that 134 million public websites are included in the Sphere dataset. In order to quickly scan millions of web citations for factual inaccuracies, it depends on the collective knowledge of the internet.

Read More

Wikipedia is the AI model’s first client, which is rather appropriate. The crowd-sourced internet encyclopedia is already utilizing Sphere to scan its pages and indicate sources that don’t genuinely support the statements in the entry, according to Meta’s release.

According to Meta, when Sphere identifies a dubious source, it will also suggest a more reliable one or a revision to help increase the entry’s accuracy.

In a statement, Meta noted that Wikipedia hosts more than 6.5 million entries in the English language alone and adds roughly 17,000 new entries to its pages each month. “Wikipedia is the default first stop in the hunt for research information, background material, or an answer to that nagging question about pop culture,” Meta said.

According to Meta, the agreement with Wikipedia does not involve cash compensation in either way. Both Meta and Wikipedia have access to enormous training grounds for Sphere, which might potentially speed up Wikipedia’s verification process and increase its factual correctness.

Automated technologies in use at the time could already recognize information that had no citations. However, the intricacy of identifying specific claims with dubious sources and establishing whether those sources genuinely support the assertions in issue “needs an AI system’s depth of comprehension and analysis,” according to Meta’s researchers.

The work Sphere did with Wikipedia, according to Shani Evenstein Sigalov, a researcher at Tel Aviv University and vice chair of the Wikimedia Foundation Board of Trustees, is “a compelling example of machine learning technologies that may assist scale the work of volunteers.”

Sigalov stated that by enhancing these procedures, “we will be able to draw new contributors to Wikipedia and offer better, more trustworthy information to billions of people around the world.”

The launch of Sphere represents Meta’s most recent effort to combat internet misinformation, and it may serve to deflect criticism of the corporation for its own part in allowing such misinformation to persist.

Users and authorities have frequently criticized Meta over the past few years for the propagation of false material on the firm’s social media platforms, which include Facebook, Instagram, and WhatsApp. Allegations that the firm prioritized profitability over combating disinformation have been strengthened by former workers and released in internal documents, and Meta CEO Mark Zuckerberg has been invited to testify before Congress about the issue.

President Joe Biden charged the social media behemoth with “killing lives” by enabling the dissemination of false information about the Covid-19 vaccination on its platforms last summer. The business retaliated, asserting that billions of users were receiving “authoritative information regarding COVID-19 and vaccines” through Facebook and Instagram.

This is all about Meta’s AI Can Scan Wikipedia, for more informative content visit