Facebook has blamed the coronavirus for hampering efforts to remove suicide and self-injury posts from its platforms.
The social network revealed it took action on significantly less material containing such content between April and June because fewer reviewers were in action as the pandemic struck.
Facebook sent its moderators home in March to prevent the spread of the virus but boss Mark Zuckerberg warned enforcement requiring human intervention could be hit.
The firm says it has since brought “many reviewers back online from home” and, where it is safe, a “smaller number into the office”.
Facebook’s latest community standards report shows that 911,000 pieces of content related to suicide and self-injury underwent action within the three-month period, versus 1.7 million pieces looked at in the previous quarter.
Meanwhile on Instagram, steps were taken against 275,000 posts compared with 1.3 million before.
Action on media featuring child nudity and sexual exploitation also fell on Instagram, from one million posts to 479,400.
Facebook estimates that less than 0.05% of views were of content that violated its standards against suicide and self-injury.
“Today’s report shows the impact of Covid-19 on our content moderation and demonstrates that, while our technology for identifying and removing violating content is improving, there will continue to be areas where we rely on people to both review content and train our technology,” the company said.
“With fewer content reviewers, we took action on fewer pieces of content on both Facebook and Instagram for suicide and self-injury, and child nudity and sexual exploitation on Instagram.
“Despite these decreases, we prioritised and took action on the most harmful content within these categories.
“Our focus remains on finding and removing this content while increasing reviewer capacity as quickly and as safely as possible.”
The tech giant’s sixth report does suggest the automated technology is working to remove other violating posts, such as hate speech, which went from 9.6 million on Facebook in the last quarter to 22.5 million now.
Much of that material, 94.5%, was detected by artificial intelligence before a user had a chance to report it.
Proactive detection for hate speech on Instagram increased from 45% to 84%.
The data also suggests improvements on terrorism content, with action against 8.7 million pieces on Facebook this time compared with 6.3 million before – only 0.4% of this was reported by a user, while the vast bulk was picked up and removed automatically by the firm’s detection systems.
“We’ve made progress in combating hate on our apps, but we know we have more to do to ensure everyone feels comfortable using our services,” Facebook said.
“That’s why we’ve established new inclusive teams and task forces including – the Instagram Equity Team and the Facebook Inclusive Product Council – to help us build products that are deliberately fair and inclusive and why we are launching a Diversity Advisory Council that will provide input based on lived experience on a variety of topics and issues.
“We’re also updating our policies to more specifically account for certain kinds of kinds of implicit hate speech, such as content depicting blackface, or stereotypes about Jewish people controlling the world.”