Google Fights Back Against 'Fake News' Appearing In Search Results

 

“Fake news” has become a catch-all phrase that Google’s VP of Engineering, Ben Gomes, interprets as “blatantly misleading, low quality, offensive or downright false information.” Web content that falls into these categories continually contaminates search results, news feeds and ad placements. Yesterday Google announced changes to Search designed to combat the appearance of fake news in search results.

The changes to Search fall under three headings: Search rankings, direct feedback tools, and enhanced transparency about how Search works. Google has taken two approaches to improving search rankings so that low-quality, fake news sites have a lower probability of appearing in search results. The most straightforward approach is adjusting their search algorithms to discount sites that contain objectionable content. Google has remained silent, as expected, on what adjustments were made.

 Google’s second approach toward improving rankings takes the form of updating their Search Quality Rater Guidelines with more detailed examples of the low quality webpages that Google wants to discount. One of the ways Google examines the quality of Search is by having humans critically evaluate real search results. The Search Quality Rater Guidelines instruct the human evaluators on how to go about their business. The Guidelines are a detailed, comprehensive and valuable resource for anyone who wants to increase their sophistication in evaluating the pages they find on the web. Check it out.

Direct feedback links are being added to Search’s Autocomplete and Featured Snippet functions to enable users to easily flag content they find objectionable. Autocomplete automatically completes search strings after the user has entered a few characters. Featured Snippets are the content previews Search displays at the top of the search results.

A link will appear below the content boxes for these functions which pops up a new box that allows users to mark objectionable returns. Users can also indicate the reason they found the content objectionable. Google plans to use this feedback to improve the algorithms that power Autocomplete and Featured Snippet. The direct feedback links were not present at the time this was written.

Google’s approach to improving transparency involves updating the content policies for Autocomplete along with updating their How Search Works webpage. The content polices for Autocomplete explain the categories Google uses to remove predictions from search strings. The webpage includes a link people can use to report a prediction they see as falling into one of the banned categories.

Website owners can access Google’s How Search Works page to gain insight into making sure their webpages don’t inadvertently fall afoul of Google’s ranking policies.

Fake news isn’t going to disappear until the people who create and spread it come to the conclusion that it’s no longer producing the desired effect. Sadly, hateful and offensive speech may never disappear. Unwanted exposure to this garbage can be decreased through steps like those Google is taking to eliminate it from Search. Whether Google’s approach will be successful rests in part on the willingness of the rest of us to provide the feedback Google is requesting.

Credit: Kevin Murnane (Forbes)