Content ModerationReplaces Human LaborVerified

Amazon Rekognition is an AWS service that other businesses license to automatically scan images and videos for inappropriate content — including explicit material, violence, hate symbols, drugs, and gambling. It allows platforms with large amounts of user-generated content to flag unsafe material before a human ever sees it.

Details

Amazon Rekognition can detect explicit adult or suggestive content, violence, drugs, tobacco, alcohol, hate symbols, gambling, and disturbing content in images and videos, returning each label with a confidence score. By using Rekognition, human moderators can review a much smaller set of content — typically 1–5% of total volume — already flagged by machine learning, enabling focus on more valuable activities at a fraction of existing cost. Customers include Dream11 (100 million users), SmugMug/Flickr, and CoStar Group. The service is a commercial API sold to third-party developers and platforms.

Products affected

Amazon Rekognition

Sources & Evidence

Company Disclosure

Company DisclosureVerified

Amazon Rekognition Content Moderation

AWS·Jan 2026

https://aws.amazon.com/rekognition/content-moderation/

Company DisclosureVerified

Improving Content Moderation with Amazon Rekognition Bulk Analysis and Custom Moderation

AWS Blogs·Apr 2024

https://aws.amazon.com/blogs/machine-learning/improving-content-moderation-with-amazon-rekognition-bulk-analysis-and-custom-moderation/

Other practices by AWS (Amazon Web Services)

OtherAWS has embedded AI agents directly into its security tooling, including a service that automatically reviews application code for vulnerabilities throughout the development process and another that accelerates incident investigation and response. These services allow security teams to scale their threat-response work without proportionally growing headcount.Data AnalysisAmazon GuardDuty is AWS's cloud security service that uses AI and machine learning to automatically monitor customer cloud environments for hacking attempts, credential theft, malware, and multi-step cyberattacks. It processes billions of signals at once and alerts security teams to critical threats without human review of every event.Data AnalysisAmazon SageMaker is AWS's unified platform for data scientists and engineers to build, train, and deploy machine learning models at enterprise scale. It allows organizations to run predictive analytics, anomaly detection, and AI-driven insights across massive datasets without managing the underlying infrastructure.

Have evidence about AWS (Amazon Web Services)'s AI practices? Submit a report.

Report a Sighting →