Belfast Telegraph

Google adds new 'upsetting-offensive' content flag to search engine guidelines

Google has added a new "upsetting-offensive" content flag to its search engine guidelines as the company makes a renewed effort to clamp down on inaccurate and offensive content.

The tech giant has given the new tool to its global review teams, known as quality raters, who analyse real search results Google sends them and are tasked with flagging potentially unsuitable content, such as pornography or so-called fake news.

Google uses the data from its quality raters to help improve its search algorithms, but the raters themselves do not have the power to alter Google search results.

The firm has not commented officially on the latest update to its search guidelines, but senior engineer Paul Haahr told industry blog Search Engine Land: "We're explicitly avoiding the term 'fake news,' because we think it is too vague.

"Demonstrably inaccurate information, however, we want to target."

In the updated guidelines, which have been publicly available since 2013, Google says the new flag should be applied to any results which promote hate or violence against a group of people based on race, religion, ethnicity, graphic violence or explicit information on how to carry out harmful activities, as well as content that would be deemed offensive or upsetting in specific regions around the world.

The guidelines for quality raters include example searches to demonstrate the feature in use, including one for "holocaust history" which shows two conflicting results referencing the atrocity.

The first result suggests the Holocaust "didn't happen" and the guidelines say this should be marked as "upsetting-offensive" by raters because many people would find such a suggestion offensive.

In contrast, a second example result from The History Channel is classed as a "factually accurate source of historical information" and would not be marked with the flag.

Flagging items does not remove or demote it in search results, but the data does help Google's machine learning better identify offensive content.

Google, along with social media giants Facebook and Twitter have been criticised in recent months over their handling of offensive and explicit content, as well as the circulation of fake news on their platforms.