A polluted information ecosystem is fed by Google News’ negligent design practices that deceive news consumers
A recent analysis by Nobl Media found that Google News failed to identify 24.3% of the opinion articles it presented in the ‘headlines’ section of the site where it mingles news and opinion.
In 2018, a Pew Research Center study showed that most Americans struggle to distinguish between fact and opinion statements. To compensate for this blind spot, news consumers often rely on labels, tags and other visual and contextual signals to help them distinguish opinion articles from factual, objective news articles.
Google News uses this technique for some — but not all — opinion and heavily biased content, according to Nobl’s analysis. They place a small, gray label following the headline, the publisher and the date published. Unfortunately, the opinion label is easy to overlook.
A Google News home page from January 13, 2020. The headlines section,
positioned at the top of the page hosts news articles as well as opinion articles.
Google News’ insufficient and inadequate use of visual signals fails to consistently provide news consumers the context they need to make informed decisions when they select and consume articles.
Nobl content quality analysis
The analysis was conducted by Nobl on a sample of 579 articles presented in the top headlines section on Google News between January 19 and January 22, 2020.
Among the opinion or heavily biased articles that were presented as news stories in the analysis were the following:
Nobl Media is a brand safety company that identifies content quality, context, and credibility to protect advertisers from toxic content. Nobl's classification engine uses machine learning combined with natural language processing on the text of each article to identify linguistic patterns that indicate sentiment, bias, credibility, and other characteristics.
Each article was designated as opinion or news by the Nobl algorithm. Articles that Nobl and Google disagreed on were reviewed by experts.
Google properly labeled 78 of 103 verified opinion articles, according to the study, but failed to identify 25 articles that were opinion or heavily biased. Those articles were visually presented as if they were legitimate, unbiased news.
The study also revealed that two of the 80 articles Google marked as opinion were not opinion.
Google News aggregates content from more than 20,000 publishers, according to Wikipedia. In December 2019, there were 470 million global visits and 211 million US visits, according to Similarweb.com, which provides website traffic analytics.
Misrepresentation leads to confusion
Important journalistic decisions made by publishers are lost in translation as Google News migrates content to its site.
Traditional media publishers use visual cues and presentation conventions to help readers distinguish objective news reporting from columns, analysis, perspective, and other forms of opinion. These signals and indicators are either absent or poorly substituted on Google News.
Moreover, if a reader selects an article with the expectation that it is news and realizes the article is biased, they may feel deceived and misplace the blame on the publisher of the article rather than the distributor, who was responsible for misrepresenting the article. This reduces trust in legitimate journalistic institutions and factual reporting.
Worse, Google News’ practices indiscriminately rewards agenda-driven publishers and objective journalists alike. This incentive system erodes the quality of online content and promotes misinformation and disinformation that is intentionally created to deceive.
Google could easily fix this
When Google launched Google News in 2002 and became an aggregator and distributor of news, they inherited responsibilities previously held by the news outlets from which they scraped the headlines. But they have not assumed those responsibilities nor have they been held accountable for those responsibilities.
Now it’s time. For the sake of all of us, Google News needs to clean up its product.
There are steps they could take that would make a difference:
1. Segregate opinion in its own section.
They already have separate sections for other categories of content (sports, entertainment, business, etc.). They should do the same for opinion and biased content (e.g. perspectives, editorials, reviews, etc.)
They should only put opinion in the opinion section and only put news in the ‘headlines’ (news) section. No more mingling the two.
Present only the best examples of news and the best examples of opinion. Articles that blur the lines between opinion and news should not be selected.
2. More accurately identify opinion and strongly biased content.
Nobl has demonstrated it is possible to use machine learning to identify linguistic patterns and classify content by its qualitative characteristics such as sentiment, bias and other factors of credibility. Google should implement the use of this technology in concert with journalists to improve their news product.
3. Provide more effective visual indicators to ensure readers know what they are about to read.
Google News needs to increase the visibility of the inconspicuous tag currently used to identify opinion, at a minimum.
They should consider the use of intuitively-colored icons next to headlines that are more salient to readers who skim and scan the page. Such a feature is available through Nobl’s browser extension (shown below) and can signal to the reader if the content is high-quality or if it warrants more critical thinking or even avoidance.
Indicators that are more visually accessible, placed in front of the headline, would make it easier for news consumers to notice when scanning the page and browsing headlines.
As long as Google News maintains the current practices, they are contributing to the mistrust of the very media sources that provide their news and enabling the growing divide in America.
—–