Science

Google offers a glimpse into its fight against fake news

Nearly 200 publishers were kicked off its advertising network in November and December of last year — some, for violating Google's new policy on serving misleading content to users.

Nearly 200 publishers were kicked off its advertising network in November and December of last year

Google says it took down 1.7 billion ads in violation of its policies in 2016. (Kim Jin-a/Associated Press)

In the waning months of 2016, two of the world's biggest tech companies decided they would do their part to curb the spread of hoaxes and misinformation on their platforms — by this point, widely referred to under the umbrella of "fake news."

Facebook and Google announced they would explicitly ban fake news publishers from using their advertising networks to make money, while Facebook later announced additional efforts to flag and fact-check suspicious news stories in users' feeds.

How successful have these efforts been? Neither company will say much — but Google, at least, has offered a glimpse.

In a report released today, Google says that its advertising team reviewed 550 sites it suspected of serving misleading content from November to December last year.

Of those 550 sites, Google took action against 340 of them for violating its advertising policies.

"When we say 'take action' that basically means, this is a site that historically was working with Google and our Adsense products to show ads, and now we're no longer allowing our ad systems to support that content," said Scott Spencer, Google's director of product management for sustainable ads in an interview.

Nearly 200 publishers — that is, the site operators themselves — were also removed from Google's AdSense network permanently, the company said.

Not all of the offenders were caught violating the company's new policy specifically addressing misrepresentation; some may have run afoul of other existing policies.

In total, Google says, it took down 1.7 billion ads in violation of its policies in 2016.

Questions remain

No additional information is contained within the report — an annual review of bad advertising practices that Google dealt with last year.

In both an interview and a followup email, Google declined to name any of the publishers that had violated its policies or been permanently removed from its network. Nor could Google say how much money it had withheld from publishers of fake news, or how much money some of its highest-grossing offenders made. 

Some fake news site operators have boasted of making thousands of dollars a month in revenue from advertising displayed on their sites.

I always say the bad guys with algorithms are going to be one step ahead of the good guys with algorithms.- Susan Bidel, senior analyst at Forrester Research

The sites reviewed by Google also represent a very brief snapshot in time — the aftermath of the U.S. presidential election — and Spencer was unable to say how previous months in the year might have compared.

"There's no way to know. We take action against sites when they're identified and they violate our policies," Spencer said. "It's not like I can really extrapolate the number."

A bigger issue

Companies such as Google are only part of the picture.

"It's the advertisers' dollars. It's their responsibility to spend it wisely," said Susan Bidel, a senior analyst at Forrester Research who recently co-wrote a report on fake news for marketers and advertisers. 

That, however, is easier said than done. Often, advertisers don't know all of the sites on which their ads run — making it difficult to weed out sites designed to serve misinformation. And even if they are able to maintain a partial list of offending sites, "there's no blacklist that's going to be able to keep up with fake news," Bidel said, when publishers can quickly create new sites.

Until advertisers have more insight into where their ads run, Bidel said, it's left to advertising platforms such as Google and Facebook to weed out offending sites.

In an email, Facebook declined to answer specific questions on its efforts — specifically, how many fake news publishers it has suspended or taken action against, the names of publishers or the amount of revenue Facebook has withheld from publishers of fake news.

Instead, the company provided a statement, attributed to an unnamed spokesperson: "It is still early days, but we're looking forward to learning and continuing to roll this out more broadly soon" — "this" referring to its previously announced tools and efforts to address fake news.

"I always say the bad guys with algorithms are going to be one step ahead of the good guys with algorithms," Bidel said. "I don't know that you're ever going to be able to eradicate this form of fraud, or any other form of fraud. But it can be brought to some acceptable level — and that level needs to be determined by the industry."

ABOUT THE AUTHOR

Matthew Braga

Senior Technology Reporter

Matthew Braga is the senior technology reporter for CBC News, where he covers stories about how data is collected, used, and shared. You can contact him via email at [email protected]. For particularly sensitive messages or documents, consider using Secure Drop, an anonymous, confidential system for sharing encrypted information with CBC News.