Last year, the FBI reported 465,676 entries for missing children in the United States.1 Many of those children are runaways—either from their homes or the care of a social services agency.
A nonprofit organization serves as a clearinghouse for critical information that can help find these children. When electronic service providers detect suspicious activity that might be a clue to locate a missing child, they pass the tip on to the organization, where analysts attempt to pinpoint a physical location for the suspected perpetrator and then deliver that information to the proper law enforcement agency as quickly as possible.
The volume of tips is massive. Last year, the organization received 8.2 million tips, which were prioritized and reviewed by a team of only 25 analysts2. Some of the tips come from identifying child pornography images that have been tagged with a technique known as hashing, which can help track illegal images as they circulate online between pedophiles.
To help solve this problem, Intel is furnishing high-power computing to help make analyzing the tips more efficient and more manageable. Bob Rogers, Intel’s chief data scientist, leads the effort.
Rogers explained how Intel uses advanced technology to help find missing and exploited children.
Bob Rogers: “We designed an enterprise data strategy that allowed them to integrate literally hundreds of different databases and data sources. We modernized their IT infrastructure with new architecture, and new hardware will help the organization process cybertips. We accelerated their ability to identify a red flag in an image.”
Question: The organization receives an incredible number of tips every day, and that volume is only expected to grow. What accounts for the growing number?
Bob Rogers: “The numbers are growing dramatically and it’s driven by these advanced technologies. Electronic service providers are using more AI to identify suspect behavior, like a chat log between an adult and a minor with explicit sexual content. There’s some AI around recognizing child pornography, although that application is still very primitive. In the future, we hope to help the organization host an innovation lab where technology companies can help create new AI tools to identify child pornography. Finally, a tech coalition of online service providers has been getting better and better at pooling their resources and getting tips to the organization. They are casting a wider net, with finer mesh, more frequently.”
“We’ve loaded hundreds of thousands of past cases that the analysts have already analyzed and we are using that as the training data to create a model that can take data from the report—think IP addresses, phone numbers, text—and determine the physical location of the suspect.”
Question: What new technology tools are in the pipeline for the organization?
Bob Rogers: “The number one goal for the organization is to get the right information to the right jurisdiction. That’s been done manually. So we’ve built a very basic machine learning model and now we are in the process of building a model that can predict what jurisdiction the report should be sent to. We’ve loaded hundreds of thousands of past cases that the analysts have already analyzed and we are using that as the training data to create a model that can take data from the report—think IP addresses, phone numbers, text—and determine the physical location of the suspect.”
Artificial intelligence is joining the fight against child exploitation not a minute too soon. According to Rogers, the organization is on track to receive 15 million tips this year.