Amid the turmoil at the Capitol, Facebook faced its own uprising

[ad_1]

WASHINGTON (AP) – As Donald Trump supporters stormed the U.S. Capitol on Jan. 6, fighting police and forcing lawmakers to hide, an uprising of various kinds is taking place inside the largest social enterprise. media around the world.

Thousands of miles away, in California, Facebook engineers are struggling to tweak internal controls to slow the spread of misinformation and exciting content. Emergency actions – some of which were reversed after the 2020 election – included banning Trump, freezing comments on groups with a note for hate speech, filtering out the “Stop the Steal” rallying cry and empowerment to content moderators to act more assertively by labeling the US a “Temporary High Risk Location” for political violence.

At the same time, frustration within Facebook exploded at what some saw as the company’s shutdown and unequal response to rising extremism in the US

“Don’t we have enough time to learn how to manage discourse without enabling violence?” an employee wrote on an internal message board at the height of the unrest on Jan. 6. “We’ve been putting out this fire for a long time and we shouldn’t be surprised that it’s out of control.”

This is a question that still rests with the company today, as Congress and regulators investigate Facebook’s part in the Jan. 6 riots.

New internal documents provided by former Facebook employee Frances Haugen provide a rare glimpse into how the company appears to be in turmoil on Jan. 6. It quickly became clear that even after years in under the microscope for inadequate oversight on its platform, the social network missed how riot participants spent several weeks vowing – on Facebook itself – to stop Congress from confirming Joe Biden’s election victory .

The documents also appear to reinforce Haugen’s claim that Facebook put growth and revenue before public safety, opening the clearest window into how Facebook’s impulses conflict – to protect its business and protect democracy – fighting in the days and weeks before the January 6 coup attempt.

This story is based in part on disclosures Haugen made to the Securities and Exchange Commission and provided to Congress in the redact form of Haugen’s legal advice. The revised versions received by Congress were obtained by an agreement of news organizations, including The Associated Press.

Facebook’s so -called “Break the Glass” emergency measures put in place on Jan. 6 are essentially a toolkit of options designed to prevent the spread of harmful or violent content that the social network first used to end the bitter elections in 2020. As many as 22 of those measures were reversed at some point after the election, according to an internal spreadsheet that analyzes the company’s response.

“When the election is over, they kill them or they change the settings before, to prioritize growth over survival,” Haugen said in an interview with “60 Minutes.”

An internal Facebook report following Jan. 6, previously reported by BuzzFeed, accused the company of having a “piece” approach to the fast -growing “Stop Stealing” pages, associated with false sources of misinformation, and violent and exciting comments.

Facebook says the situation is more complex and it is carefully calibrating its controls to respond quickly to spikes on hateful and violent content, as it did in January. The company said it was not responsible for the rioters ’actions and having tighter control over the area before that day would not help.

Facebook’s decisions to phase some safety measures in or out take into account account signals from the Facebook platform as well as information from law enforcement, spokesman Dani Lever said . “When those signals changed, the steps also changed.”

Lever said some of the measures remained in place until February and others will remain active today.

Some employees were unhappy with Facebook’s management of problematic content even before the riot on Jan. 6. An employee who left the company in 2020 left a long note charging that those promising new tools , backed by strong research, has been censored by Facebook for “public fears and stakeholder policy responses” (translation: concerns about negative reactions from Trump allies and investors).

“Similarly (albeit more with respect), I have seen the generated and functioning precautions restored for the same reasons,” the employee wrote, whose name was blacked out.

Research conducted by Facebook before the 2020 campaign left little doubt that its algorithm could pose a serious risk of spreading misinformation and potentially radical users.

A 2019 study, titled “Carol’s Journey to QAnon-A User Study of Misinfo Testing and the Risk Factors Encountered through the Recommendations System,” described the results of a experiment conducted with a test account established to demonstrate the views of a prototype “strongly conservative” – but not extremist – 41 -year -old North Carolina woman. This test account, using the fake name Carol Smith, suggests a preference for major news sources such as Fox News, followed by comedy groups that mocked liberals, embraced Christianity and was a fan of Melania Trump.

Within a day, the page recommendations for this account generated by Facebook itself changed to a “somewhat disturbing, polarizing state,” the study found. By day 2, the algorithm recommends more extremist content, including a group linked to QAnon, that the fake user not join because he or she is not instinctual with conspiracy theories.

A week later the trial subject feed featured “a barrage of intense, conspiracy and graphic content,” along with posts reviving Obama’s false lies and linking the Clintons to the murder of a former state senator of Arkansas. Much of the content was pushed by questionable groups run from overseas or by administrators with a track record for violating Facebook’s policies regarding bot activity.

The results led the researcher, whose name was changed by the whistleblower, to recommend safety measures that run from removing content with known conspiracy references and disabling “top contributors” badges for misinformation misinformation in lowering the threshold number of followers required before Facebook can verify a page administrator identity

Among other Facebook employees who read the research the response was almost universally supported.

“Hey! It’s a thorough and well-structured (and disturbing) study, “wrote one user, their name blacked out by the whistleblower.” Do you know of any concrete changes that have come out here? ”

Facebook said the study is one of many examples of its commitment to continue to analyze and improve its platform.

Another study given to congressional investigators, titled “Understanding the Risks of Dangerous Community Topics,” discussed how mentally minded people accept a topic at the borderline or Identity can generate “echo rooms” for misinformation that normalize harmful thoughts, spurs radicalization and can provide a justification for violence.

Examples of such dangerous communities include QAnon and, hate groups that promote theories of a racial war.

“The risk of offline violence or harm becomes more likely when like-minded people come together and support each other to take action,” the study concludes.

The billing documents filed by federal prosecutors against those who allegedly rushed to the Capitol have examples of similar people coming together.

Prosecutors say a senior leader of the Oath Keepers militia group used Facebook to discuss forming an “alliance” and linking plans with another extremist group, the Proud Boys, before the riot at the Capitol. .

“We decided to work together and close it,” Kelly Meggs, described by authorities as the head of the Florida chapter of Oath Keepers, said on Facebook, according to court records.

[ad_2]