Mark Zuckerberg, chief executive officer and founder of Facebook Inc., arrives for a House Financial Services Committee hearing in Washington, D.C., Oct. 23, 2019.
Andrew Harrer | Bloomberg | Getty Images
The Facebook Papers, a series of articles published by a consortium of 17 U.S. news outlets beginning on Friday, shed new light on the company’s thinking behind its actions leading up to the Capitol insurrection on Jan. 6 and its ability to fend of hate speech in languages outside of English.
Facebook shares were slightly negative in premarket trading Monday morning after the news outlets published their stories based on the leaked documents. The company is also scheduled to report quarterly earnings after markets close on Monday.
The documents were provided to the news outlets by Frances Haugen, a former Facebook employee who took tens of thousands of pages of internal research with her before she left. She’s since provided those documents to Congress and the Securities and Exchange Commission, seeking whistleblower status.
“At the heart of these stories is a premise which is false,” a Facebook spokesperson said in a statement in response to the flood of reporting. “Yes, we’re a business and we make profit, but the idea that we do so at the expense of people’s safety or wellbeing misunderstands where our own commercial interests lie. The truth is we’ve invested $13 billion and have over 40,000 people to do one job: keep people safe on Facebook.”
Here are some of the major themes the Facebook Papers have explored so far:
The documents revealed frustration among Facebook’s ranks about the company’s ability to get the spread of potential inciting content under control.
“Haven’t we had enough time to figure out how to manage discourse without enabling violence?” an employee wrote on an internal message board during the riot outside the U.S. Capitol on Jan. 6, according to the AP. “We’ve been fueling this fire for a long time and we shouldn’t be surprised it’s now out of control.”
Facebook had put additional emergency measures in place ahead of the 2020 election to stem the spread of violent or dangerous content if needed. But as many as 22 of those measures were set aside after the election and before Jan. 6, internal documents reviewed by the AP showed.
A Facebook spokesperson told the outlet its use of those measures followed signals from its own platform and law enforcement.
Some of the reports showed how Facebook’s content moderation systems can fall flat when faced with languages besides English.
The Associated Press reported that Arabic poses a particularly difficult challenge for content moderators. Arabic-speaking users have learned to use symbols or extra spaces in words thought to set off flags in Facebook’s systems, like the names of militant groups.
While the methods are meant by some to avoid an overzealous content moderation system, the AP reported that certain measures have managed to avoid Facebook’s hate speech censors.
“We were incorrectly enforcing counterterrorism content in Arabic,” an internal Facebook document said, according to the AP. Meanwhile, it said, the system “limits users from participating in political speech, impeding their right to freedom of expression.”
Facebook told the AP it’s put more resources into recruiting local dialect and topic experts and has researched ways to improve its systems.
Other reports show that some Facebook employees were dismayed by the company’s handling of misinformation in India, believing leadership made decisions to avoid angering the Indian government.
Hate speech concerns in the region were amplified by similar language barrier issues as in the Middle East. According to the AP, Facebook added hate speech classifiers in Hindi and Bengali in 2018 and 2020, respectively.
One researcher who set up an account as a user in India in 2019 found that by following Facebook’s algorithm recommendations, they saw “more images of dead people in the past three weeks than I’ve seen in my entire life total,” in the News Feed, according to The New York Times.
A Facebook spokesperson told the Times that hate speech against marginalized groups in India and elsewhere has been growing, and its “committed to updating our policies as hate speech evolves online.”
This story is developing. Check back for updates.