Skip to content
Submit incident
AII-1000-93901
ReportedReported by fewer than three independent publishers, with no official finding.Supporting sources are pending verification.

Facebook Search Algorithm Returns Adult Content to Users

Facebook's search algorithm returned adult videos to users entering common search terms, including autoplay with sound on mobile. The content evaded moderation systems using cloaking and fake metadata, representing a repeat of similar failures from the previous year.

22 June 2026 Β· Meta Β· Harmful and abusive content
22 June 2026
AII-1000-93901
Quick facts
Organization
Meta (subject)
Category
Harmful and abusive content
Country
Vietnam
Date
22 June 2026
Drafted by
Raidu Incident Detection

What happened

Facebook's search algorithm returned multiple adult videos to users searching for common keywords. When users entered popular search terms, the search results displayed explicitly sexual video content.

On mobile applications for iOS and Android, the adult content videos automatically played with sound when they appeared on screen, without requiring user interaction. The videos did not have thumbnail image modifications or masking to evade automated review systems.

Reported impact

Users encountered the inappropriate content while searching on both web and mobile platforms. On mobile, autoplay of adult content created problematic situations for users in workplaces or around children.

Index analysis

Spammers used 'cloaking' techniques combined with fake metadata descriptions to bypass Facebook's automated content moderation systems. The automated moderation system scans videos based on initial frames and random time points, which allowed obscured content to evade detection.

Governance controls

This was not the first time the search tool had this vulnerability. A similar incident occurred at the end of the previous year when users searching single characters like 'b', 'c', or 'x' received adult content in results.

Classification

Organization
Meta (subject)
Category
Harmful and abusive content
Industry
Media, social platforms and marketing
Country
Vietnam
Harm status
Actual harm
Intent
Intentional
Timing
Post-deployment
Harm types
Public interest
Harmed
General public
Failure modes
Manipulation or dark pattern, Content moderation failure
Consequences
None recorded
AI systems
Facebook Search Algorithm

Sources and evidence

Sources not yet published

This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.

How sources are verifiedSuggest a source for this incident