Tech

Google employees weighed free speech concerns before the 2016 elections, internal emails show

Key Points
  • Google and YouTube employees voiced free speech concerns before the 2016 elections, according to an internal email discussion obtained by CNBC.
  • Employees criticized an effort at crowd-sourced YouTube content moderation.
  • Workers said concern over a lack of content mediation resources was brought up in a prior all-hands meeting.
Sundar Pichai
Bloomberg | Getty Images

Google employees appear to have foreseen many of the company's political challenges in the run-up to Donald Trump's presidential election, according to an internal email discussion obtained by CNBC.

In September and October 2016, a chain of emails appeared on the Free Speech mailing list, and included messages from more than a dozen Google and YouTube employees, including former engineer and self-proclaimed conservative whistleblower Kevin Cernekee. The thread showed workers had concerns about free speech and content moderation tactics on social media sites — including its own — three years before these issues erupted into mainstream political discourse.

Government agencies and politicians from both sides of the aisle have increasingly scrutinized the company over free speech, political bias and extremist content. President Donald Trump and Republicans have attacked Google because they believe the company actively censored conservative search results and speech, most recently pointing to Cernekee's complaints as an example.

Google did not immediately respond to requests for comment.

Concerns over YouTube

In the 2016 thread, titled "More political censorship and witch hunts in tech," workers debated YouTube's efforts to curb violent content.

YouTube has been under fire for failing to moderate widespread extremism content and misinformation. YouTube also recently faced backlash for its vague policies, including when it suspended the monetization of a popular conservative user Steven Crowder hours after defending him. Soon after, the company updated its policies by banning content that displays supremacy, but critics continue asking CEO Susan Wojcicki for more specifics on moderation efforts.

In the 2016 email thread, employees discussed a company effort called YouTube Heroes, a program where YouTube community members could sign up to act as additional mediators to flag content.

One employee noted that Heroes had been publicly criticized for enabling censorship, but others disagreed, saying that Heroes was simply a way to "scale up" moderation efforts without hiring more moderators.

"At TGIF, they explicitly said the takedown call is still made by YouTube, not YouTube Heroes," one worker said, referring to a company all-hands meeting.

Another person wrote that most moderation algorithms were "not all that accurate" adding "Even if the YouTube­-employed enforcers are themselves evenhanded, they don't have the resources to police all those reports."

The debate also delved into how political philosophies at Google should inform what kind of content YouTube permits.

"From the classical liberal viewpoint, we should allow and encourage a wide range of ideas, provided they stay without some objectively defined bounds such as not making violent threats, no using obscenities or personal insults etc," one wrote.

This person contrasted that approach with a "progressive viewpoint" that believes some ideas -- such as criticizing minorities with a history of suffering oppression -- are inherently violent and should not be offered the same speech protection.

"If the progressive viewpoint does inform how we moderate things, are we being clear on that when when we communicate our policies?" this person wrote.

Attempting to draw free-speech boundaries

Employees also had a vigorous debate about the limits of free speech, which has recently become a hot topic on Capitol Hill for Republican politicians like Senator Ted Cruz. Cruz led a hearing on Capitol Hill last month where he and other judiciary committee members accused the company of suppressing conservative viewpoints in Google and YouTube search results.

"The criteria the government should be able to use to outlaw speech would be different to the criteria that YouTube should be able to use," one employee wrote in the thread.

"Free speech is a neutral concept; it protects nasty speech as well as nice," wrote one senior software engineer who worked on Google's Twitter integration, a project that decided which tweets showed up in search results. "That's kind of the whole point."

Others argued that it makes sense to define "hostile environment criteria" and ban speech that fits into it.

"I would be uncomfortable, I think, with a set of criteria under which (for instance) a group of people seriously setting out to get someone to commit suicide, by saturating their social media and other online channels with certain kinds of communications, are just exercising protected free speech that we freely allow on our platforms."

Some employees said they thought taking down the Coincidence Detector App, a Google Chrome plugin that wraps three sets of parentheses (known as "echoes") around names on web pages that appear to be Jewish, infringed on free speech.

"You think an app which adds triple ­parentheses around Jewish names somehow approaches the boundaries of free speech? Yeah, going to have to say 'not even close' there," stated one.

Most others, however, believed that a tool used to track Jewish people encouraged harassment and should not be protected.

Perhaps most notably, in a precursor to the current fierce debates over conservative censorship within the company, one wrote, "I just hope the alt-­right isn't taking an innocent concept like free speech and perverting it for their own ends."

WATCH: Trump criticizes Google CEO, says platform supported Clinton in 2016

VIDEO2:0102:01
Trump criticizes Google CEO, says platform supported Clinton in 2016