Article from CNBC Make It
Facebook says it wants to help fix misinformation spread online.
Facebook’s parent company, Meta, a few weeks ago introduced a new Artificial Intelligence (AI) tool called Sphere. They aim to help detect and address misinformation, or “fake news,” online. Meta claims it is “the first [AI] model capable of automatically scanning hundreds of thousands of quotes at once to check whether they actually support the claims.”
The announcement comes after years of criticism over Facebook’s own role in allowing online misinformation to flourish and spread rapidly across the globe. Sphere’s data includes 134 million public pages on the internet, according to Meta’s research team. It relies on the collective knowledge and insights of the internet to rapidly scan hundreds of thousands of citations across the web, looking for factual errors.
So it’s perhaps fitting that Meta is training its AI model using Wikipedia entries. According to Meta’s announcement, Sphere is already scanning pages in the crowd-sourced online encyclopedia to test its ability to flag sources that don’t actually support the claims in the entry.
Meta also says that when Sphere notices a suspicious source, it can recommend a stronger one – or a correction – to help improve the accuracy of the input.
"Wikipedia is the default first stop when looking for research information, background material, or an answer to that nagging question about pop culture," Meta said in a statement, noting that Wikipedia hosts more than 6.5 million entries in the English language alone and adds approximately 17,000 new entries to its pages each month.
The company also released a video showing how Sphere works:
A Wikipedia spokesperson tells CNBC Make It that the online encyclopedia is not officially partnering with Meta on its development of Sphere, and that none of Wikipedia’s entries are being automatically updated. Meta also told TechCrunch earlier this month that there is no financial compensation in any way.
Existing automated systems were already capable of identifying pieces of information that lacked citations. But Meta researchers say the complexity of isolating individual claims with questionable sources and determining whether those sources actually support the claims in question “requires the depth of understanding and analysis of an AI system.”
In a statement, Shani Evenstein Sigalov — a Tel Aviv University researcher and vice chair of the Wikimedia Foundation Board of Trustees — called Sphere's training with Wikipedia "a powerful example of machine learning tools that can help scale the work of volunteers."
“Improving these processes will allow us to attract new editors to Wikipedia and provide better, more reliable information to billions of people around the world,” Sigalov said.
Sphere marks Meta's latest attempt to address online misinformation — while potentially avoiding criticism over the company's own role in allowing this misinformation to persist.
Meta has faced repeated criticism over the past few years from users and regulators for spreading misinformation on the company's social media platforms, which include Facebook, Instagram and WhatsApp. Former employees and leaked internal documents have fueled claims that the company has inflated profits due to its fight against misinformation, and Meta CEO Mark Zuckerberg has been subpoenaed by Congress to discuss the issue.
Last summer, President Joe Biden accused the social media giant of “killing people” by allowing Covid-19 vaccine misinformation to spread on its platforms. The company pushed back, claiming that Facebook and Instagram were providing “authoritative information about COVID-19 and vaccines” to billions of users.