Advanced AI-based Automatic Redaction Software

redactX's technology solution under this Small Business Innovation and Research (SBIR) project addresses the Strategic Capability of “Responsibilities to Organize, Train, and Equip the Force.” Specifically, the mission need to protect court related documents, CUI, and PII by using Computer-vision/AI to rapidly and accurately identify and redact key information according to user specifications. redactX has developed key technology that automatically redacts text, faces, objects, and spoken words in images, videos, audio files, and documents based on user criteria. The proposed SBIR effort aims to address two main objectives for DAF: 1) Developing advanced named entity recognition for concept redaction in documents, and 2) Enhancing the user interface to streamline workflows as it pertains to text documents. The first objective involves extracting entities like dates, locations, and names for redaction, regardless of their format in the document. The second objective involves extending the user interface to utilize automated redaction, allowing for faster, higher-quality results. Additionally, a stretch goal is to develop a feature to confirm which document in a group of similarly named documents corresponds to the final version of that document, though the feasibility and accuracy of doing this is still being determined. Ultimately, the above development under this proposed Phase II will save the JAG Corps thousands of hours and increase information security leading to a better and more equipped force.