Tuesday, March 06, 2018

Protecting Profit Is The Fundamental Objective


RollingStone |  For years, YouTube profited off all kinds of extremist content; its three-strike policy was directed at copyright infringement. Its current and newly aggressive posture towards content stems from the advertiser revolt that erupted following Trump’s surprise victory. Within weeks of the 2016 election, brands like Johnson & Johnson, and ad-tech companies like AppNexus, began taking steps to distance themselves from Breitbart and other purveyors of "fake news" and extremist content. In early 2017, companies like Starbucks and Walmart started pulling their ads from YouTube, worried that their marketing was sandwiched between clips featuring foaming-at-the-mouth racists and child abusers. In a watershed moment, the global buying agency Havas pulled its ads from Google/YouTube U.K., after the Times of London detailed how ads for well-known charities were supporting Neo-Nazi articles and videos. When the influential research group Pivotal downgraded Google stock from a buy to a hold, Google suddenly grew concerned about the kind of content its proprietary algorithms had been promoting for years – intentionally and by design.

This is not a conspiracy theory worthy of a "strike," but the testimony of a former YouTube engineer named Guillaume Chaslot, who was profiled by the Guardian in early February. Chaslot, a Ph.D. in artificial intelligence, explained how his team at YouTube was tasked with designing algorithms that prioritized “watch time” alone. “Watch time was the priority,” he told the paper. “Everything else was considered a distraction… There are many ways YouTube can change its algorithms to suppress fake news and improve the quality and diversity of videos people see… I tried to change YouTube from the inside, but it didn’t work.” 

When Chaslot conducted an independent study of how his algorithms worked in the real world, he found that during recent elections in France, Germany and the U.K., YouTube "systematically amplifie[d] videos that are divisive, sensational and conspiratorial." (His findings can be seen at Algotransparency.org.) At the height of the advertising revolt, in March of last year, YouTube announced that it was "taking a hard look at our existing community guidelines to determine what content is allowed on the platform – not just what content can be monetized." CEO Susan Wojcicki announced the company would hire thousands of human moderators to watch and judge all content on the site.

YouTube's new policies were part of an industry-wide course correction. Over the past year, under the banner of combatting hate speech and fake news, Google and Facebook began to cut off search traffic and monetized content-creator accounts, not only to dangerous scam-artists like Jones, but to any site that garnered complaints or didn't meet newly enforced enforced and vaguely defined criteria of "credible" and "quality."