Amazon receives over 1 million AI-related child sexual abuse material reports in 2025, but refuses to disclose source of the content.
The National Center for Missing and Exploited Children (NCMEC) reported more than one million cases of AI-related child sexual abuse material (CSAM) in 2025, with Amazon being responsible for a vast majority of these reports. According to an investigation by Bloomberg, Amazon found the material in its training data but refused to disclose where it came from.
Amazon attributed the high volume of CSAM reports to the fact that it uses third-party sources to train its AI services and warned that it cannot provide further details about the source of the content due to concerns over diluting the efficacy of other reporting channels. This means that Amazon is unable to pass on any actionable information to law enforcement.
The NCMEC's CyberTipline, where companies are legally required to report suspected CSAM, received a high volume of AI-related reports from various companies in 2025, but with one notable exception: Amazon's reports proved to be "inactionable" due to the lack of source information.
Amazon has since taken steps to improve its detection and removal of CSAM content. However, this raises concerns over how the company is handling sensitive data. The NCMEC is urging AI companies to take greater responsibility for preventing CSAM on their platforms.
The growing issue of CSAM in AI chatbots and models is becoming increasingly alarming, with several high-profile cases involving teenagers who took their own lives after being exposed to abusive content generated by these platforms. Meta, OpenAI, and Character.AI are among the companies facing lawsuits related to these incidents.
In response to the rising concerns, Amazon emphasized its commitment to preventing CSAM across all of its businesses. However, experts are warning that more needs to be done to ensure AI models are designed with robust safeguards to prevent the spread of abusive content.
The National Center for Missing and Exploited Children (NCMEC) reported more than one million cases of AI-related child sexual abuse material (CSAM) in 2025, with Amazon being responsible for a vast majority of these reports. According to an investigation by Bloomberg, Amazon found the material in its training data but refused to disclose where it came from.
Amazon attributed the high volume of CSAM reports to the fact that it uses third-party sources to train its AI services and warned that it cannot provide further details about the source of the content due to concerns over diluting the efficacy of other reporting channels. This means that Amazon is unable to pass on any actionable information to law enforcement.
The NCMEC's CyberTipline, where companies are legally required to report suspected CSAM, received a high volume of AI-related reports from various companies in 2025, but with one notable exception: Amazon's reports proved to be "inactionable" due to the lack of source information.
Amazon has since taken steps to improve its detection and removal of CSAM content. However, this raises concerns over how the company is handling sensitive data. The NCMEC is urging AI companies to take greater responsibility for preventing CSAM on their platforms.
The growing issue of CSAM in AI chatbots and models is becoming increasingly alarming, with several high-profile cases involving teenagers who took their own lives after being exposed to abusive content generated by these platforms. Meta, OpenAI, and Character.AI are among the companies facing lawsuits related to these incidents.
In response to the rising concerns, Amazon emphasized its commitment to preventing CSAM across all of its businesses. However, experts are warning that more needs to be done to ensure AI models are designed with robust safeguards to prevent the spread of abusive content.
amazon's lack of transparency on this whole thing is just mind-blowing... like what even is the point of having a third-party report system if you're not gonna give any details? 
and meanwhile, companies are all like "oh no, our AI is bad" but really it's just a case of being lazy and not wanting to deal with the messy details 
anyway, glad they're taking steps to improve detection (i guess) but i'm still waiting for some real accountability on this one 
like what?! 1 million reports of child sexual abuse material on Amazon? That's just crazy 
and I'm not surprised they won't tell us where it came from, that sounds super dodgy 
. But at the same time, I get why they're being all vague about it, I mean we don't want to compromise the whole reporting system, right? 
And what's up with all these AI chatbots and models that can generate super bad content? It's like, how do we even regulate this stuff? 
. We need to do something to prevent this from happening more often, you know? Like, we need to make sure AI companies are taking responsibility for what they're creating and how it affects people 
.
this is getting out of hand, like literally 1 million reports and amazon just says it's a third-party thing 
and yeah, the ncmeC is right on point too... these companies need to step up their game and design better safeguards for their models 
It's wild that they won't disclose where it came from, but at the same time, I get why they can't - it could mess with other reporting channels and stuff. 
It's crazy to think that these platforms can be created to spread such hate and abuse... I mean, we're basically giving them permission to do so. 
... I mean, a million reports of child abuse material from AI is just too much. Amazon's not being transparent enough about where it came from tho 
It's not like they're getting these reports for free, right? 
It's like they're saying "we found this stuff, but you'll never know where it came from" 
... that's a good start. But we need to talk about how these companies are handling sensitive data in the first place 
. They can't just keep it on the down-low and expect everything to be okay 
.
What if those companies aren't even reporting everything they find? Or what if the AI is picking up on stuff it shouldn't be seeing in the first place? 
It's like, Amazon's trying to protect itself from getting overwhelmed, but at the same time... shouldn't they be taking responsibility for letting this stuff go through their systems in the first place? 
