The Center for Data Innovation spoke with Richard Brown, a technology advancement officer at the International Center for Missing and Exploited Children and coordinator of Project Vic, a collaborative effort between law enforcement agencies to algorithmically identify never-before-seen child pornography images seized in police raids and help investigators rapidly determine if they represent new cases of child exploitation. Brown discussed how the technology works and why phone camera metadata has been one of investigators’ greatest assets.
This interview has been lightly edited.
Travis Korte: Can you introduce Project Vic and describe what impact you hope it will have on law enforcement efforts to reduce child exploitation?
Richard Brown: Project Vic is a grassroots effort by law enforcement agencies to standardize the way certain law enforcement tools can interact with each other. In particular, we think investigators could benefit from aggregate data produced by all the different agencies that fight child exploitation crimes. We knew law enforcement agencies were working in silos and seeing the same child pornography images over and over again, and if we combined our efforts and started sharing cooperatively we knew it could save investigators extra time. With that extra time they can go and look through never-been-seen images to make sure there are no victims out there that we could be rescuing.
TK: How do you know if an image has been seen before or not?
RB: Once an investigator reviews a seizure of child pornography images the investigator will scale and rate those images into categories and send them into our cloud-based hashing service. When new investigators go out into the field to do seizures, they pull down the most up-to-date hash list, load it into their triage tools, and that can help them identify images in new seizures that have already been seen from ones that have never been seen.
TK: What changes in the investigation once you find something new?
RB: Investigators can look into it right away. We can go through hundreds of thousands of images quite quickly. In a collection of thousands of images, we might find 500 never-before-seen images and we want investigators to look very close at those to see who is in the background—if it’s family members, say—as well as what kind of camera is being used and other information. And then we want to do everything possible to take the children in those images out of harm’s way.
TK: Can you talk about some of the successes that have come from the project so far?
RB: We talked to investigators about what kind of information they wanted in the hash and a lot of them stressed how much metadata iPhones and Android phones can attach to an image, including camera information, latitude and longitude, serial number, etc. Traditionally law enforcement has stripped a lot of that information out but that doesn’t make sense because it could lead to the location where the image was taken. So we’ve had cases where that information has lead to specific locations or specific phones being used in the household. And investigators are able to track down the offender and say, “this is your phone’s serial number.” There are also people who have traveled around to different states committing these crimes and you can prove that in the metadata.
TK: What’s the scope of the project now, and what’s your outlook for the future?
RB: The project’s been going on for two years. The first step was to get agencies to clean out their closets of images and collate them into one spot where we can de-duplicate and start categorizing those images. The project was started with three or four local task forces along with a handful of federal agencies. Today there are over 50 agencies, both domestic and international, that are involved in some way with the project. They have benefitted both from workload reduction as well as minimizing the front line investigators’s exposure of child pornography. It used to take weeks to get a full investigation of a hard drive, and we’re tightening that up a lot. In the last six months it’s really caught on and really grown exponentially in the United States and abroad. Project Vic has been very aggressive, and we’ve had very good responses from big corporations that are willing to donate technology when they see the good work being done. And the federal government is now looking at this project as a viable one to get behind.