Google’s fake news crackdown: Search demotes shady results, adds user feedback

Google wants to improve search quality by surfacing authoritative news and offering better feedback mechanisms to report examples of fake or harmful news.

googlesearchautocomplete.jpg 

Google is adding a new ‘Report inappropriate predictions’ link in Search to let users provide feedback on Autocomplete suggestions.

Image: Google

 

 

Google is stepping up its war on fake news by tweaking search to demote low-quality content and offering tools for users to report offensive Autocomplete and Featured Snippets suggestions.

The move addresses criticisms that the search giant and other social platforms haven’t done enough to combat the spread of fake news, an issue that came into focus during the recent US elections.

The company already has measures in place to deal with search spam and other deceptive practices.

But until now it hasn’t offered a comprehensive answer to its challenges with fake news, such as the recent finding that a top result on Google search for “Did the Holocaust happen” linked to a neo-Nazi site. Autocomplete meanwhile was found suggesting, “Are Jews evil?” after typing in “Are Jews…”.

Google admitted on Tuesday that more “structural changes” in Search are needed to address these problems and outlined a number of ways it is handling the issues.

The company notes that the fake news problem is fairly limited in search, with 0.25 percent of queries returning offensive or clearly misleading content.

However, as Search Engineland points out, because the numbers are low, Google was caught off guard by a PR crisis that emerged in recent months over shoddy results. A steady stream of examples began to cast doubt on the quality of Google search.

To address this perception, Google is changing its algorithms to “surface more authoritative content” and last month updated its Search Quality Rater Guidelines with clear examples of misleading information, unexpected offensive results, and hoaxes and conspiracy theories.

Google’s human reviewers, equipped with the new guidelines, then rate the quality of results as it makes changes to its algorithm, which will help its algorithms further demote low-quality content.

It’s also introducing direct feedback tools for Autocomplete suggestions and Featured Snippets boxes. Last month Google was called out for pulling a number of fake news sources in Featured Snippets, including a result from a conspiracy site that claimed former president Barack Obama was planning a “communist coup”.

“The content that appears in these features is generated algorithmically and is a reflection of what people are searching for and what’s available on the web. This can sometimes lead to results that are unexpected, inaccurate or offensive,” said Ben Gomes, Google’s vice president of engineering.

Autocomplete now features a ‘Report inappropriate suggestions’ link below the search box, allowing users to tag if they encounter a suggestion that’s hateful, sexually explicit, violent, or harmful.

The existing feedback option for Featured Snippets has been updated to let users report examples of bogus, hateful, and vulgar information in the boxes. Google intends to use this feedback to improve its algorithms.

No Comments Yet.

Leave a comment