Amazon's Reporting of Child Sexual Abuse Material in AI Training Data

In 2025, the National Center for Missing and Exploited Children (NCMEC) received over 1 million reports of AI-related child sexual abuse material (CSAM), with the vast majority originating from Amazon.com Inc. The company reported hundreds of thousands of suspected CSAM pieces found in its AI training data, raising significant concerns among child safety officials regarding the lack of transparency about the origins of this material. Amazon's reports to NCMEC have been characterized as “inactionable” due to insufficient details about the sources of the content, which could hinder law enforcement efforts to identify and prosecute offenders.

Amazon Found ‘High Volume’ Of Child Sex Abuse Material in AI Training DataAmazon Discovers Hundreds of Thousands of Child Sexual Abuse Images in AI Training DataAmazon discovered a 'high volume' of CSAM in its AI training data but isn't saying where it came from

Amazon's reporting practices have come under scrutiny as the company has stated that it obtained the inappropriate content from external sources used to train its AI services. According to Fallon McNulty, executive director of NCMEC’s CyberTipline, the absence of source information makes it impossible for NCMEC to trace the material's origin or take further action. McNulty noted that while other companies provided actionable data, Amazon's reports lacked the necessary details, rendering them ineffective for law enforcement.

Amazon Discovers Hundreds of Thousands of Child Sexual Abuse Images in AI Training Data

Amazon has defended its practices, asserting that it employs an over-inclusive threshold for scanning training data, which results in a high percentage of false positives. The company emphasized its commitment to responsible AI and stated, “While our proactive safeguards cannot provide the same detail in NCMEC reports as consumer-facing tools, we stand by our commitment to responsible AI and will continue our work to prevent CSAM.” However, critics argue that the lack of transparency in reporting undermines efforts to combat child exploitation.

Amazon discovered a 'high volume' of CSAM in its AI training data but isn't saying where it came fromAmazon Discovers Hundreds of Thousands of Child Sexual Abuse Images in AI Training DataAmazon Found ‘High Volume’ Of Child Sex Abuse Material in AI Training Data

Child safety advocates have expressed alarm over the implications of Amazon's reporting gap. Experts warn that the proliferation of AI-generated CSAM, which saw a staggering increase in reports from 4,700 in 2023 to over 67,000 in 2024, poses a significant threat to child safety. Kerry Smith, Chief Executive of the Internet Watch Foundation, highlighted the ease with which offenders can now create abusive material using AI tools, stating, “Criminals essentially can have their own child sexual abuse machines to make whatever they want to see.”

Amazon Found Child Abuse Material in AI Training Data but Withheld Key Details

While Amazon reported a dramatic increase in CSAM findings, other tech companies like OpenAI and Google have also faced challenges in detecting and reporting similar content. OpenAI reported over 75,000 depictions of child exploitation in the first half of 2025, indicating a broader industry issue. However, the disparity in the volume and detail of reports raises questions about the effectiveness of current detection methods and the potential for contaminated training datasets.

Amazon Found Child Abuse Material in AI Training Data but Withheld Key Details

As regulatory scrutiny increases, California Attorney General Rob Bonta has opened an investigation into xAI and Grok regarding their image generation capabilities and potential links to child sexual abuse material. This investigation reflects a growing concern about the adequacy of safeguards implemented by AI companies to prevent the creation of illegal content. The ongoing challenges in addressing the surge of technology-facilitated child abuse underscore the urgent need for improved transparency and accountability in AI training practices.

Amazon Found Child Abuse Material in AI Training Data but Withheld Key Details