On Thursday, Meta apologised for an error that led to a surge of graphic and violent content appearing on Instagram Reels.
This apology came after numerous users reported encountering disturbing videos, including those depicting violence and gore, despite having their “Sensitive Content Control” settings enabled to the highest moderation level.
The error and its impact
According to a Meta spokesperson, the error caused some Instagram users to see content in their Reels feed that should not have been recommended.
“We have fixed an error that caused some users to see content in their Instagram Reels feed that should not have been recommended. We apologise for the mistake,” the spokesperson stated.
This incident raised concerns about potential issues with Instagram’s algorithm and content moderation system.
Users worldwide took to social media platforms to express their unease, with some questioning how such material bypassed Meta’s safeguards.
Instagram users reported seeing a flood of short-form videos showing gore and violence, including killings and cartel violence. These videos were marked with the “sensitive content” label but were being recommended back-to-back. Despite Meta’s policy to remove particularly violent or graphic content and add warning labels to others, the technical error bypassed these measures.
Read also: Google, Meta, X face sanctions over anti-competitive practices against South African media platforms
Meta’s response and policy
Meta’s content policies typically remove “particularly violent or graphic” material and add warning labels to other sensitive content.
Access to such content is also restricted to users under 18.
However, Meta allows some graphic content if it helps users condemn and raise awareness about important issues like human rights abuses or armed conflicts, provided it comes with limitations such as warning labels.
The incident follows recent changes to Meta’s content moderation practices, including replacing third-party fact-checkers with a community-driven flagging system and simplifying content policies. Despite these changes, the error highlighted ongoing challenges in maintaining effective content moderation on the platform.