REDACT: PII Redaction and Privacy Protection with Ollama Integration

Protection of Personally Identifiable Information (PII) in government documents is required for privacy and data security. Traditional redaction of PII is slow, time-consuming, and imprecise for bulk use. In this paper, an automated and secure PII redaction system with support for various document types, including PDFs and scanned images, is presented. Through the application of Optical Character Recognition (OCR) in extracting text and Large Language Models (LLMs) through Ollama in PII identification, our solution ensures accurate detection of sensitive information in various languages. Users have access to review and customize redaction rules in order to create flexibility and autonomy in the process. With the addition of advanced document processing techniques and visual redaction for image- based data, the solution suggested enhances security, ease of use, and conformance with data protection law.