Brand Safety Policy

False Positive & Brand Impersonation Handling Guidelines

Version: 1.0
Last Updated: October 2025
Contact: support@rakshalink.in

1. Purpose

This policy outlines RakshaLink's approach to responsibly handling false positives, brand impersonation detections, and communications from verified companies or institutions whose legitimate messages may have been incorrectly flagged as suspicious.

RakshaLink is committed to ensuring user safety without compromising brand integrity or spreading misinformation.

2. Core Principles

a. Advisory Nature of Detection

RakshaLink's scam detection is probability-based and advisory, not definitive.
All messages are evaluated using machine learning models that assign confidence scores (e.g., "We're 85% sure this looks suspicious") rather than absolute classifications.

RakshaLink does not claim that any specific company or brand is fraudulent.
We detect patterns resembling impersonation or risky behavior.

b. Brand Impersonation Focus

RakshaLink's detections are aimed exclusively at messages impersonating or spoofing well-known institutions.

Legitimate Entity Impersonation Example
Bank bank_kyc.apk (fake app)
Government Portal myAadhar-upd.apk (clone)
E-Commerce Site offer_claim_2025.apk (fake giveaway)

c. On-Device Analysis Only

All scam detection occurs locally on the user's device.
RakshaLink does not upload, transmit, or store any user messages, sender details, or brand content to the cloud.

This ensures:

  • Full compliance with GDPR, CCPA, and India's DPDP Act
  • Zero third-party data processing
  • No external liability from message content

3. False Positive Mitigation

RakshaLink minimizes false positives using:

  1. Dynamic Whitelist — Verified sender IDs (e.g., VM-SBIUPI, AD-AADHAR)
  2. Official Domain Verification — Prioritizes .gov.in, .org.in, .bank, etc.
  3. Multi-Signal Detection — Scams trigger alerts only when multiple risk patterns align
  4. User Overrides — Users can mark detections as "Safe" for on-device model refinement

4. Response to Brand Notifications

When a legitimate brand contacts RakshaLink about false positives:

  1. Acknowledge Promptly
    "Thank you for bringing this to our attention. RakshaLink uses automated on-device analysis and does not label legitimate brands as fraudulent. The detection likely flagged an impersonation pattern. We will review and update our whitelist immediately."
  2. Verify Sender Patterns
    Review logs, rules, and sample messages.
  3. Update Whitelist / Rules
    Modify patterns to prevent future false positives.
  4. Confirm Resolution
    Notify the brand after correction.
  5. Record Update
    Log changes under Brand False Positive Handling.

5. Disclaimer of Liability

RakshaLink provides scam detection insights based on automated pattern recognition. While every effort is made to ensure accuracy, false positives or negatives may occur.

RakshaLink and its developers are not responsible for any direct or indirect damages arising from interpretation of detection results.

Users are advised to verify suspicious messages directly with official sources before taking action.

6. Transparency to Users

RakshaLink informs users that:

  • Detection results are probabilistic, not absolute
  • Verified sources should always be cross-checked
  • Trusted senders can be marked "Safe"

7. Periodic Review

This policy will be reviewed bi-annually or after major feature updates affecting detection logic, brand handling, or privacy compliance.

Contact Us

For any concerns related to brand representation or false detections:

📩 Email: support@rakshalink.in
🌐 Website: https://rakshalink.in

Document Owner: Mohan Gangahanumaiah
App: RakshaLink — Scam Detection & Privacy-First Safety
© 2025 RakshaLink