The tech industry’s crackdown on content from right-wing users spread this week from social media platforms to behind-the-scenes companies that do not typically take on free speech issues, a sign of heightened aggressiveness ahead of a planned far-right gathering this weekend inspired by the riots in Charlottesville.
Microsoft, which provides cloud services that host websites, on Thursday threatened to cease hosting an alt-right-leaning social network called Gab.ai after a user there posted comments that threatened Jewish people with “ritual death by torture.” The content, Microsoft said, “incites violence, is not protected by the First Amendment” and violated the company’s own policies.
Microsoft then issued an ultimatum: Gab must remove the posts or Microsoft would terminate the site’s access to its cloud service, Azure. Gab ultimately complied. The demand appeared to mark the latest escalation by tech giants to strike the right balance between facilitating free speech and combating hateful, abusive or violent content online.
"We believe we have an important responsibility to ensure that our services are not abused by people and groups seeking to incite violence against others,” Microsoft said.
Gab did not respond to an email seeking comment. Andrew Torba, the founder of Gab, wrote in a post that the social network is seeking a new hosting provider and considering how to build “our own infrastructure.”
“We believe this was the best decision for the longevity of the platform and the war against Silicon Valley,” he wrote.
The threat of censorship by a cloud provider is much more unusual than the removals by sites such as Facebook, Twitter and YouTube of content that violates their policies. One such high-profile case came this week, when Facebook, Apple, YouTube and others deleted years of content from conservative conspiracy site Infowars over allegations of hate speech. Cloud-hosting services, which are less visible to the user, tend to take a more neutral stance on a site’s content than platforms such as Facebook and YouTube, which have lengthy policies that outline acceptable behavior and teams of moderators on staff to review it.
For experts, the question is whether all technology companies have the same expectations and responsibilities. There can be inconsistencies even within the same company. For instance, YouTube belongs to Google — but Google Play’s app store let the Infowars app stay up, where it surged in popularity.
Google is also one of the main cloud providers, along with Microsoft and Amazon.com. It is unclear whether Google or Amazon’s cloud hosts content that has been banned by social media. (Amazon.com founder and chief executive Jeffrey P. Bezos owns The Washington Post.)
As neo-Nazi rallies in Charlottesville demonstrated last year, tech platforms can serve as both organizing tools and catalysts for violence. Alt-right groups initially tried to use Facebook to coordinate their efforts at the Unite the Right rally, and after clashes in Virginia concluded, sites such as the Daily Stormer praised the death of Heather Heyer, a counterprotester who was killed at the rally. In response, tech companies faced immense pressure to rethink their policies and take down the offending content. Facebook removed organizers' posts while infrastructure services such as Cloudflare and GoDaddy blocked the Daily Stormer.
It is harder to make the argument of responsibility to prohibit content for companies that are deeper in the tech infrastructure, said Kevin Bankston, director of the Open Technology Institute at the New America think tank.
Right-leaning groups have called some of these removals a form of censorship, and their attacks on technology companies have become an election-year political rallying cry. That tension resurfaced this week around Infowars. Some prominent conservatives, including President Trump’s son Donald Trump Jr., quickly attacked tech companies for stifling right-leaning views.
Like its peers, Microsoft Azure provides the technical architecture to host websites generally without regard to their topic or purpose — so long as their makers are not engaged in illegal activity, for example, or intending to “violate the rights of others,” according to the company’s “acceptable use policy.” Other major cloud providers, including Amazon and Google, maintain similar rules on what is considered appropriate to save on their own servers.
To that end, Microsoft appeared to contact Gab on Aug. 9 to inform it of a policy violation for “malicious activity,” according to an image that Gab tweeted Thursday. A user had flagged the comments on Gab, according to Microsoft. Microsoft said the complaint was inadvertently mis-tagged as a type of malicious activity called phishing.
In response, Gab wrote to Microsoft — another email it posted online — to take issue with its rationale. By Thursday evening, though, the user behind the offensive posts, @Patrick_little, had not taken them down, so Gab removed them. Some Gab users responded by blasting Microsoft’s demand as a form of censorship.
Bankston, who said Microsoft acted appropriately, said Web-hosting companies should be treated more like a phone provider — which would not be held responsible if someone discussed committing a terrorist or violent act on their services — rather than like a social network, which has a community of users it holds to a set of values and standards. He also said the companies need to be transparent about why they take content down, pointing out that they get into trouble when their decisions appear to be opaque.
“In a lot of ways, the alt-right is in a win-win situation, and the platforms are in a lose-lose,” Bankston said. “Either the platforms do take their content down and they claim martyrdom, or they don’t take their content down and their messages continue to get out.”