A Citizen-Centric Anonymous Whistleblower Reporting System Using NLP and Secure Tracking

A Citizen-Centric Anonymous Whistleblower Reporting System Using NLP and Secure Tracking

Veeksha A

Student, St Joseph Engineering College, Mangalore

Ms Priyadarshini P

Assistant Professor, St Joseph Engineering College, Mangalore

Abstract: Whistleblowing systems are critical tools in combating corruption, fostering ethical conduct, and promoting transparency. However, widespread underreporting persists due to fears of retaliation and a lack of anonymity. This research presents a citizen-centric anonymous whistleblower reporting system that enables secure, anonymous, and streamlined complaint reporting.

Our platform eliminates the need for user registration or identification and assigns a unique tracking code for status monitoring. It integrates Natural Language Processing (NLP) to analyse complaint narratives and categorize issues by type and priority. A secure admin portal with data visualizations facilitates timely and informed decision-making. This paper evaluates the systems usability, security, and effectiveness, showing high user satisfaction, robust anonymity, and improved responses to high-priority cases, offering a practical model for future anonymous reporting systems.

Index Terms: Anonymous reporting, Natural Language Processing (NLP), secure complaint system, whistleblower protection.

1. INTRODUCTION

   Whistleblowing systems have become vital tools for promoting transparency, accountability, and ethical governance. Yet, even with technological advancements and increased awareness, many individuals remain hesitant to report wrongdoing. The fear of retaliation and the lack of guaranteed anonymity often prevent people from speaking up, leaving critical issues unreported and unresolved.

   To address these concerns, this research presents a citizen- focused anonymous whistleblower reporting platform designed to ensure complete privacy, usability, and efficiency. The system allows users to report complaints without logging in or providing personal information. Upon submission, each complaint is assigned a unique tracking code, enabling users to follow up on their report without compromising their identity.

   At the core of the platform is the integration of Natural Language Processing (NLP), which helps in analyzing and

   categorizing complaints based on their content and urgency. This approach not only simplifies complaint handling but also ensures that high-priority cases receive timely attention. By automating classification, the system minimizes the manual workload for administrators and enhances response accuracy.

   To support decision-making and improve administrative workflows, the platform includes a visual dashboard. This feature allows authorized personnel to view complaint trends, filter cases by category, and track resolution progress through clear visualizations and analytics. Security is reinforced through encrypted data storage and restricted access, ensuring the confidentiality of all submissions.

   This system aims to create a safe and accessible space for individuals to report issues without fear. By blending automation, privacy, and transparency, the platform offers a practical and scalable model for anonymous reporting across various sectors such as public services, education, and corporate environments. The results demonstrate that when trust and technology are combined thoughtfully, citizen engagement and institutional accountability can be significantly strengthened.

2. LITERATURE SURVEY

   Natkar et al.[1]: This study explores the development of a secure, anonymous whistleblower system emphasizing non- identifiable complaint submission. The proposed system eliminates login requirements and uses AES-based encryption to store complaints safely. A major contribution is the generation of a unique tracking ID to follow the complaint status anonymously. The authors advocate for a minimalistic and intuitive user interface to enhance accessibility for citizens, including those with limited technical literacy. Their evaluation demonstrates that complaint submission rates improve significantly when anonymity and ease-of-use are prioritized in the system design.

   Jobinpicard and Doha[2]: This paper outlines critical system design objectives for whistleblower platforms using a design science research methodology. The authors identify key pillars such as trust preservation, content integrity, and fair

   stakeholder representation. A dual-layer architecture is proposed, where user identity and complaint data are processed separately to ensure privacy. The study also emphasizes the importance of feedback loops, secure message handling, and transparency in complaint processing. Their recommendations provide an ethical and systematic blueprint for developing reliable whistleblowing frameworks, especially in organizational settings.

   Young and Farshadkhah[3]: This research addresses a key weakness in anonymous systems: the challenge of complaint credibility. The authors propose utilizing Self-Sovereign Identity (SSI) and blockchain technology to enable whistleblowers to submit verifiable credentials (e.g., proof of employment) without disclosing their identity. The SSI system enables organizations to prioritize complaints from users with authentic affiliations while maintaining full anonymity. Their implementation model also includes zero- knowledge proof mechanisms to validate user claims securely. The paper presents a breakthrough in balancing trustworthiness and confidentiality in whistleblower systems.

   Pramono and Aruzzi[4]: This study evaluates the implementation of whistleblowing systems in Indonesian public institutions. The authors highlight significant policy and technical shortcomings, such as a lack of encryption, poor follow-up mechanisms, and public mistrust. Their findings suggest that many complaints are underreported due to fear of retaliation and lack of transparency in the resolution process. The authors recommend stronger legal frameworks, institutional training, and awareness campaigns to improve adoption and system effectiveness. Their analysis underscores the importance of combining technological innovation with supportive governance to ensure real-world impact.

3. METHODOLOGY

   The proposed Anonymous Whistleblower Reporting System follows a structured multi-step architecture designed to ensure secure, anonymous, and efficient handling of complaints. The system integrates modern web technologies, NLP techniques, and secure tracking mechanisms to protect user identity and streamline administrative response.

   1. Complaint Submission

      The system begins with a user-friendly, web-based interface where individuals can submit their complaints without requiring a login or personal information. The form includes Complaint type (e.g., corruption, harassment, misuse of power), detailed description (text input), and optional file upload. This anonymous interface empowers users to report misconduct without fear of retaliation or exposure of their identity.

   2. Preprocessing

      Once the complaint is submitted, the text undergoes preprocessing before NLP analysis, is the removal of stop words, special characters, and extra whitespace, tokenization, and lowercasing, normalization of spelling and formatting. This step ensures clean and uniform input for NLP classification, improving both speed and accuracy.

   3. NLP-Based Classification and Prioritization

      The system utilizes a Natural Language Processing (NLP) engine to analyze the complaint text. The process includes keyword extraction, categorization into types (e.g., Fraud, Workplace Abuse, Ethical Violation, etc.), and priority scoring. This utomated classification helps reduce administrative burden and ensures serious complaints are prioritized appropriately.

   4. Track Code Generation

      After classification, the system generates a unique, anonymous tracking code for the complaint as a secure alphanumeric hash is created without linking to any user ID. This tracking code allows users to check the complaint status later. No login or personal verification is required. This step balances user anonymity with transparency and follow-up capability.

   5. Admin Dashboard and Review Panel

      The backend admin panel is designed to help reviewers manage complaints efficiently. Complaints can be filtered by category, priority, and charts and graphs display trends and complaint volumes. Admins can view attached files, download reports, and update statuses. This enhances response time and allows data-driven decision-making.

   6. Secure Storage and Encryption

      All complaint data is stored securely in a MongoDB database with encryption at rest for all text and file data, a hash-based ID for complaint records, and access restricted to verified admin roles. This system ensures that no identity information is ever stored, maintaining full user anonymity.

   7. Status Tracking by Users

      Users can visit the tracking portal, enter their unique code, and check the complaint status (Received / Under Review / Resolved). This provides reassurance to the whistleblower and encourages continued engagement with the system.

      Fig.1. Workflow Diagram

4. RESULTS AND ANALYSIS

   The proposed anonymous whistleblower reporting system was tested in a simulated environment using synthetic and real complaint data. Its performance was evaluated based on classification accuracy, user anonymity retention, system responsiveness, and admin-side usability.

   • NLP Classification Module (Accuracy:91.4%) Classification Capability: The NLP engine accurately categorized complaints into predefined categories such as Corruption, Abuse, Service Delay, and Ethical Violations.

     Strengths: The model used a combination of rule-based keyword matching and machine learning classifiers to maintain high accuracy, even with unstructured or informal text.

     Implications: With over 91% classification accuracy, the system significantly reduces manual sorting burden and ensures high-priority cases receive faster administrative attention.

   • Anonymous Tracking System

     Tracking Capability: The secure hash-based tracking system successfully generated unique tracking codes in 96.5% of test submissions. The remaining cases were impacted by user-side issues such as page reloads or code loss.

     Strengths: No identity is required or stored, and code collisions were not observed in test batches.

     Limitations: Users must securely store their tracking code, as no reset or recovery is possible due to enforced anonymity.

     Usefulness: Enables efficient, private follow-up without

     compromising whistleblower identity.

   • Admin Dashboard Usability

   Admin Feedback: In mock user testing with five admin users, the dashboard interface was rated 4.6/5 for clarity, ease of use, and utility.

   Visualization Capability: The dashboard provided filters for category, urgency, and date range, along with pie charts and trend graphs using Chart.js.

   Impact: Admins were able to process and assign complaints 35% faster compared to traditional spreadsheet-based methods.

5. FUTURE ENHANCEMENTS

   To further enhance the effectiveness, security, and scalability of the anonymous whistleblower reporting system, the following future improvements can be considered:

   • Biometric-Backed Admin Access: Incorporate biometric authentication (such as fingerprint or facial recognition) for admin login to ensure only authorized personnel can access sensitive complaint data. This adds an extra layer of protection against unauthorized access or account misuse, improving trust and internal accountability.

   • Blockchain Integration: Leverage blockchain technology to create an immutable ledger of complaint records and actions taken. This will improve transparency, provide tamper-proof audit trails, and increase the overall integrity of the complaint-handling process.

   • Sentiment and Tone Analysis: Integrate advanced sentiment analysis into the NLP engine to evaluate the emotional tone of complaints (e.g., urgency, distress). This will enable the system to prioritize cases not only based on keywords but also emotional intensity, helping admins respond more effectively.

   • Mobile App Interface: Develop a dedicated mobile application for users to submit complaints anonymously, upload voice messages or photos, and track status using a tracking code. This improves accessibility and encourages broader public participation, especially in rural or mobile-first communities.

   • Multi-Language NLP Support: Expand NLP capabilities to support multiple regional languages. This ensures inclusivity, allowing people from diverse linguistic backgrounds to comfortably submit complaints without language barriers.

   • Anonymous Feedback Loop: Introduce a feature where users can receive follow-up updates or feedback messages through the tracking code without revealing their identity. This creates a sense of engagement and builds trust in the platform's responsiveness.

6. CONCLUSION

   This system addresses critical limitations in traditional whistleblower platforms by offering an end-to-end secure, anonymous, and intelligent complaint submission and tracking mechanism. By integrating NLP-based classification, a secure tracking engine, and a robust admin interface, it empowers institutions to handle citizen concerns more efficiently while protecting the whistleblowers identity. This approach is scalable, modular, and easily adaptable across multiple domains, including education, public services, and corporate governance.

7. REFERENCES

1. S. R. Natkar, M. D. Patil, and S. M. Sayyad, Secure and Anonymous Whistleblower System with Tracking ID, International Journal of Innovative Science and Research Technology, vol. 5, no. 6, pp. 567 572, 2020.

2. J. Jobinpicard and D. Doha, Design Considerations for Digital Whistleblowing Systems: Trust, Privacy, and Transparency, in Proc. 15th Int. Conf. Information Systems Security and Privacy (ICISSP), pp. 121129, 2021.

3. S. Young and A. Farshadkhah, Anonymous Yet Credible Whistleblower Systems Using SSI and Blockchain, IEEE Access, vol. 9, pp. 9432094331, 2021.

4. A. Pramono and F. Aruzzi, Challenges in Implementing Whistleblower Systems in Government Institutions, in Proc. Int. Conf. E-Governance, vol. 12, no. 2, pp. 4552, 2020.

5. D. Jurafsky and J. H. Martin, Speech and Language Processing, 3rd ed., Upper Saddle River, NJ, USA: Pearson, 2023.

6. M. Sahlgren and A. Cöster, Using Bag-of-Concepts to Improve the Performance of Support Vector Machines in Text Classification, in Proc. 20th Int. Conf. Computational Linguistics (COLING), pp. 487 493, 2004.

7. N. Kshetri, Privacy and Security Issues in e-Government, Government Information Quarterly, vol. 32, no. 3, pp. 258268, July 2015.

8. K. Sampigethaya and R. Poovendran, Privacy and Security in Online Government: Issues and Challenges, IEEE Security & Privacy, vol. 7, no. 3, pp. 3037, MayJune 2009.