Google AI engineers have created a machine learning model that identifies consensus on controversial topics, the company has announced. The system is being used by Google to prevent misinformation from appearing in the “featured snippets” that appear at the top of its search results.
Google parses relevant sources using its “unified multitasking model” (MUM), a “multimodal” natural language processing technology it developed last year, to reduce the risk of misinformation appearing in featured snippets.
“Our systems can now understand the notion of consensus, which is when several high-quality sources on the web agree on the same fact,” Pandu Nayak, Google’s vice president of search, wrote in a statement. blog entry about the MUM.
Featured snippet content can now be compared to “high-quality web sources to see if there is a general consensus for that mention, even if the sources use different words or concepts to describe the same thing,” according to Nayak.
“We found that using this consensus-based technique significantly improved the quality and usefulness of featured snippet mentions.”
The new system also reduces the chance of misleading snippets appearing in response to queries based on incorrect assumptions. The most recent update uses MUM to identify such search queries, reducing the probability of incorrect snippets appearing by 40%.
Nayak outlined some additional measures to help promote “information literacy” among its users. These include the new “About This Result” feature, which pulls information from Wikipedia about linked sources from search results, and “Content Notices,” which notify users when Google is unsure of the quality of the information of its results.
The updates are part of Google’s concerted effort to combat misinformation on its platforms. “Google was founded on the premise that information can be a powerful tool for people around the world,” Nayak wrote.
“We’re determined to continue to do our part to help people around the world find what they’re looking for and provide the context they need to make informed decisions about what they watch online.”
Other measures include a $75 million investment from the Google News Initiative to help develop media literacy and a new project to develop information literacy lesson plans for school children, according to Nayak.
How to apply for the 2022 Innovation Awards?
Eddie is an Australian news reporter with over 9 years in the industry and has published on Forbes and tech crunch.