Meta Apologizes After Technical Error Floods Instagram Reels with Violent Content

Meta Apologizes After Technical Error Floods Instagram Reels with Violent Content

Meta issued an apology on Thursday after a technical error caused Instagram users worldwide to encounter graphic violence, gore, and explicit content in their Reels feeds, even when strict content filters were enabled.

Why it matters:

The incident has raised significant concerns about Meta’s content moderation systems, especially after the company recently announced major changes to its approach, including ending third-party fact-checking in favor of a community-driven model similar to X’s.

Technical Failure:

The error impacted Instagram’s recommendation algorithm in several troubling ways. Users reported encountering extremely disturbing content that violated Meta’s own policies on graphic material, including:

• Videos showing people being killed or severely injured
• Content labeled as “sensitive” appearing in regular feeds
• Material that should have been restricted appearing for all users

User Experience:

The disturbing content appeared suddenly and without warning for many Instagram users. Several reports across social media platforms revealed that users were subjected to disturbing videos in feeds that typically contained benign content.

One Reddit user described seeing “street fights, school shootings, murder, and gruesome accidents,” while another reported a drastic shift in their feed from “planes, watches, painting, and cats” to “body horror and videos with Russian descriptions.” Some users even encountered child exploitation material.

“We have fixed an error that caused some users to see content in their Instagram Reels feed that should not have been recommended,” a Meta spokesperson said, according to CBS. “We apologize for the mistake.”

The company has yet to explain what caused the error or provide specific details on how many users were affected. Meta’s policies prohibit the display of extremely graphic content, including videos showing dismemberment, visible entrails, or charred bodies, as well as material featuring sadistic remarks on human or animal suffering.

A Growing Concern:

This incident comes on the heels of significant changes at Meta, including the elimination of its third-party fact-checking program in January and substantial workforce reductions. The company has laid off approximately 21,000 employees over 2022 and 2023, many from its trust and safety divisions.

Meta has explicitly denied any connection between this error and its recent policy changes regarding fact-checking. However, some users and industry observers have speculated that the reduction in human oversight might have contributed to the failure.

Despite the controversy, Meta’s stock rose by approximately 1.5% in premarket trading on Thursday, continuing a trend that has seen shares increase nearly 40% over the past year.