You need to agree to share your contact information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

Log in or Sign Up to review the conditions and access this dataset content.

👉 Looking for the open multilingual baseline? Start with ai4privacy/pii-masking-openpii-1.5m (1.5M samples, 30 languages, open-PII taxonomy).

🇪🇺🌏 Personal Health & Medical Information, Global PII Dataset

👁 PII-Masking-3M Coverage

Part of PII-Masking-3M by Ai4Privacy, the global (2M base + Asia Pacific) PII-masking corpus.

📖 More information: www.ai4privacy.com/datasets/pii-masking-3m-asia-pacific

---

PII Label Distribution

👁 Bar chart showing PII label distribution across entity types

Full label inventory:

---

Coverage

New Asia Pacific locales (added in 3M): Vietnamese (vi-VN), Indonesian (id-ID), Malay (ms-MY), Filipino (tl-PH), Chinese (zh-CN), Japanese (ja-JP), Korean (ko-KR), plus English variants in Singapore (en-SG) and India (en-IN).

---

Language Distribution

👁 Bar chart showing entry distribution across 30 languages

---

Data Format

{
 "source_text": "Original text with synthetic PII values",
 "masked_text": "Text with [LABEL_N] placeholders",
 "privacy_mask": [{"label": "GIVENNAME", "start": 0, "end": 5, "value": "Alice"}],
 "uid": 12345,
 "language": "de", "region": "DE", "script": "Latn",
 "mbert_tokens": ["Alice", "hat"],
 "mbert_token_classes": ["B-GIVENNAME", "O"]
}

---

PII-Masking-3M Collection

---

Usage

from datasets import load_dataset
dataset = load_dataset("ai4privacy/pii-masking-health-phi-400k")

Commercial Licensing

This is an enterprise dataset under the PII-Masking-3M programme. Contact Ai4Privacy for licensing terms covering production use.

---

Access & Partnerships

• 🔓 Request data access: Open Data Access Form
• 🤝 Commercial partnerships: Enterprise Contact Form

---

p5y Data Analytics

This dataset is built on the p5y framework, think of it as i18n but for privacy. Just as i18n (internationalization) translates content into different locales, p5y translates sensitive data into privacy-safe formats through a standardized 3-step approach:

1. Awareness, Scan and markup private entities in unstructured text, producing a structured privacy mask with entity types, distribution, density, and risk assessment.
2. Protection, Control identified personal data through masking, pseudonymization, or k-anonymization, tailored to the specific use case and regulatory requirements.
3. Quality Assurance, Measure remaining privacy risk after anonymization, evaluating de-anonymization risks through expert annotation and automated assessment.

Learn more at p5y.org

---

Legal Disclaimer

The dataset is provided "as is" without any guarantees or warranties, express or implied. Ai4Privacy and Ai Suisse SA make no representations regarding accuracy, completeness, or suitability for any specific purpose. Users utilize the dataset at their own risk and bear full responsibility for any outcomes. Under no circumstances shall Ai4Privacy, Ai Suisse SA, or affiliates be held liable for any damages arising from use of the dataset. Users are responsible for ensuring compliance with all applicable laws, regulations, and ethical guidelines, including GDPR, CCPA, Singapore PDPA, Japan APPI, Korea PIPA, Indonesia UU PDP, and AI-related legislation.

---

ai4privacy.com · Discord · Ai Suisse SA

PII-Masking-3M, the world's largest open multilingual PII masking corpus, covering Europe, Americas, and Asia Pacific across 30 languages. • 11 items • Updated 16 days ago