A recent investigation by The Guardian has revealed that social media companies relying on artificial intelligence (AI) software to moderate their platforms are generating unviable reports on cases of child sexual abuse, hindering US police investigations and delaying the identification of potential predators. These companies are required by law to report any child sexual abuse material detected on their platforms to the National Center for Missing & Exploited Children (NCMEC), which then forwards the leads to relevant law enforcement agencies.
In 2022, NCMEC received over 32 million reports of suspected child sexual exploitation, with Meta being the largest reporter, responsible for over 27 million reports. However, the use of AI-generated reports poses a significant challenge for law enforcement. Due to privacy protections under the fourth amendment, officers and NCMEC are not permitted to open reports without a search warrant unless a human at the social media company has reviewed the contents. This legal requirement has resulted in delays of several weeks, allowing potential evidence to be lost and offenders to remain undetected for longer periods.
Child safety experts and attorneys have expressed concern about the detrimental impact of these delays on community safety. Investigations are stalled, and law enforcement is unable to take immediate action to protect potential victims. In some cases, social media companies even disable user accounts after submitting a report, potentially leading to the removal of crucial evidence from their servers.
AI-generated tips are often not investigated because they lack the specific information needed to obtain a probable cause affidavit for a search warrant. This places an additional burden on already overwhelmed law enforcement agencies, who do not have the necessary resources to prioritize these tips.
The reliance on AI for content moderation is seen by experts as an insufficient solution. While AI can help alleviate the workload of law enforcement officers, it cannot replace the need for human intervention. As a Massachusetts-based prosecutor pointed out, AI may not be able to identify new child abuse material and can only analyze existing data points. The importance of human review cannot be understated when dealing with these sensitive cases.
The current situation has prompted concerns about the efficacy of AI moderation and the need for increased resources and support for law enforcement agencies to tackle the growing issue of child sexual abuse online.
The source of the article is from the blog agogs.sk