META Meta Platforms Inc

By Niharika Mandhana and Rhiannon Hoyle 

Facebook Inc. left a gruesome video of shootings at a New Zealand mosque up on its site for half an hour after a user brought it to the company's attention, a previously unreported delay that highlights the social media company's difficulty in controlling violent content on its platform.

The timetable, based on times supplied overnight by Facebook and the New Zealand police, means the video showing men, women and children being gunned down at a mosque in Christchurch was available on Facebook's site for an hour from the time it started rolling to the time it was ultimately removed.

The live-streaming video was taken by the shooter as he began a rampage that left 50 people dead Friday afternoon. The broadcast began on Facebook at around 1:33 p.m. and ended 17 minutes later at approximately 1:50 p.m. A Facebook user flagged the post to the company 12 minutes after that, Facebook said. The social media team of the New Zealand police alerted Facebook to the video at 2:29 p.m., police told the Journal. Facebook said it took the video down within minutes thereafter.

New Zealand police said the timeline was approximately correct. Facebook declined to comment.

The delay was the result of a string of gaps in systems Facebook has set up to stop violent videos and other objectionable content from appearing on its platform. Artificial intelligence programs the company relies upon to block bad content failed to catch the video during its live broadcast and afterward. Then the alert from the user, who flagged the video after the broadcast ended, wasn't fast-tracked, because Facebook prioritizes warnings about suicide attempts in such cases.

By the time the video was taken down, it had been viewed 4,000 times on the site, Facebook said. It was copied millions of times on Facebook and other sites beyond Facebook's control.

Hany Farid, a professor of computer science at Dartmouth College, said the company for years focused on expanding aggressively and unleashing potent new products without putting the necessary protections in place beforehand.

"The problem is that there are just no safeguards," Mr. Farid said. "The company can't say these are unintended or unforeseen consequences. These were foreseeable."

Facebook has said it is doing more to moderate the large numbers of posts that go online. It has doubled down on AI as a solution, in addition to using some 15,000 human content reviewers. In a post on the company's website Wednesday, Guy Rosen, Facebook's vice president for integrity, acknowledged limitations in the company's handling of live broadcasts and said the company's artificial intelligence tools hadn't been able to catch the video.

"We recognize that the immediacy of Facebook Live brings unique challenges," he wrote in his post on the company's website.

The killings last week left New Zealand reeling. Police have charged Australian Brenton Tarrant with murder in the case, saying he mowed down worshipers who had gathered for Friday prayers at two mosques. He hasn't entered a plea.

The social media giant has come under fire in many parts of the world--from Myanmar to Sri Lanka--for failing to take sufficient action against hate speech, fake news and misinformation on its site. The problems are amplified with Facebook Live, the video broadcast tool it introduced in 2016. Facebook said users post millions of live videos a day. In the past, they have included live beatings and murders.

Australian Prime Minister Scott Morrison asked over the weekend whether Facebook should be allowed to offer services such as live videos if it can't control them.

Some critics have called on the site to impose a time delay during which videos could be checked. Mr. Rosen said a delay would be swamped by the high volume of live videos broadcast daily. He also said it would slow reports by users that could enable authorities to provide help on the ground.

Facebook has touted the success of its technical tools in tackling some kinds of terrorist content. The company said last year that Facebook itself--not its users--was identifying nearly all of the material related to Islamic State and al Qaeda that it removed. But the platform, accessed each month by some 2 billion users posting in more than 80 languages, hasn't had that success with other kinds of extremist content.

In his post, Mr. Rosen said artificial intelligence has worked well to identify nudity and terrorist propaganda, because there are lots of examples of such content to train the machines. But the AI didn't have enough material to learn to recognize content like the mosque massacre, he said, because "these events are thankfully rare."

Once the video was flagged by a user, it appeared to get stuck in queue to be reviewed. Mr. Rosen said the social media site accelerates its review of all live videos that are flagged by users while they are still being broadcast. The policy also applies to videos that have recently ended, but only in the case of suicide reports. The Christchurch video wasn't reported as a suicide and as a result wasn't prioritized.

The company will re-examine its procedures and expand the categories for accelerated review, Mr. Rosen said.

Mr. Farid, the Dartmouth professor, said the delay in taking the video down even after it was flagged by a user shows Facebook's legions of content checkers aren't enough for a platform of Facebook's size.

Some of New Zealand's largest advertisers, including ASB Bank, suspended advertising on Facebook and Google after the shooter's video was posted.

Lindsay Mouat, the CEO of the Association of New Zealand Advertisers, said New Zealand corporations "are just appalled with the way that live video streaming of what happened could actually get up and occur in Facebook Live without any moderation."

Facebook's systems caught 1.2 million videos of the attack as they were being uploaded in the first 24 hours; another 300,000 made it to the site and were removed afterward. But users were able to make copies of the video on other sites to keep it alive.

Mr. Rosen said a user of the website 8chan, which is often home to white supremacist and anti-Muslim content, posted a link to a copy of the video on a file-sharing site from where it spread more broadly.

Facebook struggled to keep the video from reappearing on its own site, he said, because a "core community of bad actors" worked to continually re-upload edited versions to defeat the company's detection tools. Facebook found and blocked more than 800 variants of the video that were circulating, he said.

Write to Niharika Mandhana at niharika.mandhana@wsj.com and Rhiannon Hoyle at rhiannon.hoyle@wsj.com

(END) Dow Jones Newswires

March 21, 2019 11:00 ET (15:00 GMT)

Copyright (c) 2019 Dow Jones & Company, Inc.