How the FBI plans to manage crowdsourced video
Connecting state and local government leaders
The Federal Bureau of Investigation is looking for a citizen-facing system to help with ingesting, cataloging and routing image and video content submitted by the public.
In high-profile criminal investigations, justice officials frequently ask the public to submit images and video that may help solve crimes. But these crowdsourced submissions can easily overcome the storage and processing capabilities of police and justice departments.
In an effort to address this challenge, the Federal Bureau of Investigation is looking for a citizen-facing system to help with ingesting, cataloging and routing image and video content submitted consensually by the public.
According to the FBI’s recent request for information, the system should have the following capabilities:
Secure evidence handling: The product should ensure that unmodified original submissions are available and that the integrity of evidence is never compromised.
User-friendly upload experience. The product should allow expert and novice users to upload materials without forcing them to transcode or modify their media before submission. Additionally, the product must make clear to users the legal ramifications of uploading data to the FBI and its intended use of that data.
Security, availability and scalability. The product must scale to accommodate massive collections while maintaining uninterrupted availability and must be secured from those who may want to disable the system or alter submissions.
Integration with state and local authorities. A solution should be available to, or compatible with, state and local collection systems, while allowing the FBI to obtain the original data and annotations made by the original collector.
Automated tagging. The product should tag and filter certain information about the media, such as conflicts in the EXIF, camera identification, computer vision tools and more.
Triage/analyst tools. The product should accommodate large groups of users simultaneously without a significant formal training burden.
Interchangeable components. The product may be expected to pass information submitted by the user via another product to any outputs downstream, via APIs, for example.
Responses are due by Nov. 17, 2014.