SAN FRANCISCO (Reuters) – Google said on Monday it is working on a policy change to prevent websites that misrepresent content from using its AdSense advertising network, a move aimed at halting the spread of “fake news” and other types of misinformation on the internet.
The shift comes as Google, Facebook and Twitter face a backlash over the role they played in the US presidential election by allowing the spread of false and often malicious information that might have swayed voters toward Republican candidate Donald Trump.
The issue has provoked a fierce debate within Facebook especially, with Chief Executive Mark Zuckerberg insisting twice in recent days that the site had no role in influencing the election.
Google’s move does not address the issue of fake news or hoaxes appearing in Google search results. That happened in the last few days, when a search for “final election count” for a time took users to a fake news story saying Trump won the popular vote. Votes are still being counted, with Democratic candidate Hillary Clinton showing a slight lead.
Nor does it suggest that the company has moved to a mechanism for rating the accuracy of particular articles.
Rather, the change is aimed at assuring that publishers on the network are legitimate and eliminating the financial incentives that appear to have driven the production of much fake news.
“Moving forward, we will restrict ad serving on pages that misrepresent, misstate, or conceal information about the publisher, the publisher’s content, or the primary purpose of the web property,” Google said in a statement.
The company did not detail how it would implement or enforce the new policy.
Tiny publishers in Macedonia
AdSense, which allows advertisers to place text ads on the millions of websites that are part of Google’s network, is a major source of money for many publishers.
A report in BuzzFeed News last month showed how tiny publishers in Macedonia were creating websites with fake news much of it denigrating Clinton which were widely shared on Facebook.
That sharing in turn led people to click on links which brought them to the Macedonian websites, which could then make money on the traffic via AdSense.
Facebook has been widely blamed for allowing the spread of online misinformation, most of it pro-Trump, but Zuckerberg has rejected the notion that Facebook influenced the outcome of the election or that fake news is a major problem on the service.
“Of all the content on Facebook, more than 99 percent of what people see is authentic,” he wrote in a blog post on Saturday. “Only a very small amount is fake news and hoaxes.”
Google has long had rules for its AdSense program, barring ads from appearing next to pornography or violent content. Work on the policy update announced on Monday began before the election, a Google spokeswoman said.
The company uses a combination of humans and artificial intelligence to review sites that apply to be a part of AdSense, and sites continue to be monitored after they are accepted, a former Google employee who worked on ad systems said. Google’s artificial intelligence systems learn from sites that have been removed from the program, speeding the removal of similar sites.
The issue of fake news is critical for Google from a business standpoint, as many advertisers do not want their brands to be touted alongside dubious content. Google must constantly hone its systems to try to stay one step ahead of unscrupulous publishers, the former employee said.
Google has not said whether it believes its search algorithms, or its separate system for ranking results in the Google News service, also need to be modified to cope with the fake news issue.
Fil Menczer, a professor of informatics and computing at Indiana University who has studied the spread of misinformation on social media, said Google’s move with AdSense was a positive step.
“One of the incentives for a good portion of fake news is money,” he said. “This could cut the income that creates the incentive to create the fake news sites.”
However, he cautioned that detecting fake news sites was not easy. “What if it is a site with some real information and some fake news? It requires specialized knowledge and having humans [do it] doesn’t scale,” he said.
(Reporting by Julia Love; Editing by Dan Grebler, Jonathan Weber and Bill Rigby)