Advocates fault Facebook over misleading posts by politicos

Some civil rights groups agree Facebook has made progress in addressing their concerns, but advocates are still criticizing the tech giant's reiteration that it won't remove newsworthy posts from elected officials, even if they're misleading or break the site's rules against things like racist speech.

The criticism came Thursday at a meeting in Atlanta hosted by Facebook and one of its most persistent critics, the civil rights group Color of Change. The meeting was aimed at raising outstanding concerns and addressing solutions as Facebook works to complete a civil rights audit due in early 2020.

Facebook Chief Operating Officer Sheryl Sandberg told the roughly 100 invited guests that her company doesn't want to shy away from talking about "how our platform still harms people and allows harm to happen," after the company came in for a storm of criticism following revelations about how the social media site was used to spread disinformation during the 2016 election.

"We know better than most companies, that we have a lot to do in terms of strong actions to restore confidence," Sandberg told attendees at the end of the day.

The Menlo Park, California, company didn't make any promises in Thursday's discussions, although executives said they had clearly heard the discontent over the newsworthiness rule.

Sandberg and other company executives emphasized that the newsworthiness policy is not a blanket exemption for politicians to engage in hate speech and also that the company is still considering how it will handle the matter going forward.

It's not a new policy for Facebook, but drew wide notice earlier this week when British politician-turned Facebook executive Nick Clegg discussed the issue in a speech Tuesday in Washington.

"The understanding that we are exempting all political speech from all politicians, that's just not accurate," Facebook Public Policy Director Neil Potts told The Associated Press.

Facebook agreed to the civil rights audit after years of criticism and moved forward with Thursday's meeting following long negotiations.

"Racism is not like the problems Silicon Valley is used to solving," said Color of Change President Rashad Robinson. "You just can't hack it. You just can't assign a project team to figure it out. Racism is highly adaptable. You block it one area and it goes and attacks in another."

A number of those who attended emphasized that it's important that Facebook entrench a focus on civil rights.

"I think it's essential that Facebook institutionalize this kind of input, and that it understands that civil rights issues will be a part of the company's concerns and problems for its entire existence," said Laura Murphy, who's leading the office.

Sandberg, who jotted notes in a notebook while on stage and in the audience, said the company is paying attention.

"We at Facebook recognize that our platforms are part of this new front and this new battleground," she told The Associated Press.

But she warned that the company and critics were unlikely to agree on everything.

The newsworthiness policy highlights a key disagreement. Advocates say they too often see racism and white supremacy thriving, while those who try to talk frankly about racism in an effort to fight it end up in Facebook "jail," with posts removed or accounts suspended. Henry Fernandez said the newsworthiness exception is "bad policy, made in fear that conservatives will say Facebook is silencing Trump."

"This rhetoric, whoever it comes from, and certainly if it comes from the president of the United States, it should not be on Facebook," Fernandez said. "It is killing people.... It is not news that Donald Trump says racist things. And because politicians have a larger influence in moving public opinion, they should be held to a higher standard, not a lesser one."

Since the audit began, Facebook has banned messages discouraging people from voting and announced it would protect against interference in the census. The company in March banned explicit praise of white nationalism and separatism, although a June update on the audit said Facebook hasn't done enough to combat content that supports those ideologies without explicitly using those terms.

The company also says that its artificial intelligence systems are doing better at recognizing hate speech, detecting 65% of hate speech that was eventually removed before a user reported it in March, up from 24% in December 2017.

On the same day that Facebook announced the civil rights audit, it also announced a review led by Jon Kyl, an Arizona Republican and former U.S. senator, of claims that Facebook was biased against conservatives. That document, released in August, made no specific findings. It did catalog conservative fears about Facebook, including concerns that changes in how posts appear in newsfeeds disadvantaged right-leaning media and that Facebook's employees are too liberal. The 133 conservatives interviewed also expressed concerns that hate speech policies would work against conservatives.

Supporters of the civil rights audit slammed the conservative review, saying it set up a false equivalency between the two groups' concerns.

© 2019 The Associated Press. All rights reserved.