News Brief

Google Details its Battle Against Disinformation

At the Munich Security Conference last week Google unveiled new policies to tackle disinformation across its search engine, YouTube, Google News and advertising products at a time when the tech company is under attack for making fake news popular and easy to access.

The 30-page report gives the most detailed explanation to date about how Google defines disinformation — “deliberate efforts to deceive and mislead using the speed, scale, and technologies of the open web” — — as well as how its algorithms find and destroy such content.

The company asserts its search algorithms do not personalize content based opinions, issues or political leanings. Instead, the company says that user results, especially via its search engine, rely more on location or search language than the users’ search history.

The report also attempts to address criticism against YouTube that has grown due in part to a new study that shows the ease at which conspiracy theories are distributed and promoted, such as easily disprovable opinions like the Earth is flat. Or when users are bombarded with misleading videos about prominent individuals as what happened, The Washington Post reported, last year when users searched the initials “RBG” for information on Supreme Court Justice Ruth Bader Ginsburg when she fell ill.

Such criticism prompted Google to launch a worldwide effort to repair the credibility of the company’s platforms. In July, YouTube’s chief product officer and chief business officer announced a $25 million investment. Those funds are part of the Google News Initiative – a $300 million commitment to “strengthen quality journalism, to empower news organizations through technological innovation” and facilitate their growth.

Google’s new report outlines actions that users can make when they detect deceptive content ranging from fake news stories to spam, hate speech or harassment.

It also says that it is working with new algorithms as well as human reviewers to remove posts that violate its Community Guidelines.