What is sensitive data?
Sensitive data is any information that could cause harm to a person or an organization if it is exposed, changed or misused. It includes personal data such as Social Security numbers and health records, payment card data, passwords and API keys, and confidential business information such as deal terms, salary details and legal strategy.
For example, a sales leader asks an AI assistant to draft a renewal proposal. The assistant pulls in the customer’s discount, the contract terms and an internal note that flags the account as a churn risk. None of that is personal data, yet if the discount reached another customer or the churn note reached the client, the damage would be real.
Types of sensitive data
| Category | Examples |
|---|---|
| Personally identifiable information (PII) | Names, Social Security numbers, passport numbers, home addresses |
| Protected health information (PHI) | Diagnoses, medical record numbers, claims linked to a patient |
| Payment card data (PCI) | Card numbers, expiration dates, security codes |
| Special category data | Racial or ethnic origin, religious beliefs, genetic and biometric data, sexual orientation |
| Credentials and secrets | Passwords, API keys, access tokens |
| Business-sensitive information | Discount details, deal terms, HR cases, confidential leadership communications, legal strategy |
Sensitive data under privacy laws
Privacy laws give some categories extra protection:
- GDPR special categories (Article 9): racial or ethnic origin, political opinions, religious or philosophical beliefs, trade union membership, genetic data, biometric data used to identify a person, health data, and data about sex life or sexual orientation. Processing them is prohibited unless a specific exception applies.
- CCPA/CPRA sensitive personal information: includes Social Security and driver’s license numbers, account log-ins with passwords, precise geolocation, racial or ethnic origin, health data and the contents of private messages. Consumers can limit how businesses use it.
Business-sensitive information is mostly outside these laws, but contracts, trade-secret rules and internal policy still govern it.
Why sensitive data is harder to control in AI systems
Pattern-based tools are good at finding a Social Security number or a card number, since those values have a fixed shape. They struggle with sensitive data that depends on context. A sentence like “we will go to 30% off if they push back” contains no identifier, yet it is one of the most sensitive lines in a sales thread. AI systems also move data across tools quickly: a single AI agent may read a file, query a database and post a summary in a few seconds. Knowing what is sensitive, and where it goes, is the first step. See this primer on sensitive data discovery.
How Protecto helps
Protecto detects 200+ entity types across PII, PHI and PCI, plus custom entities, and also handles context-based sensitive data such as risk analysis, deal terms, HR matters and legal strategy. It shows where sensitive data flows and reports on how data policies are enforced. Learn more about data leak prevention for AI.
Frequently asked questions
What is considered sensitive data?
Any data that could harm a person or organization if exposed. Common examples are PII, health records, payment card data, credentials and confidential business information such as pricing and legal plans.
Is sensitive data the same as PII?
No. PII is one type of sensitive data. Sensitive data also covers health and payment data, secrets such as API keys and business information that identifies no one.
What is sensitive personal data under the GDPR?
The GDPR calls it special category data. It covers racial or ethnic origin, political opinions, religious or philosophical beliefs, trade union membership, genetic and biometric data, health data, and sex life or sexual orientation.
Related terms
Personally identifiable information (PII) · Protected health information (PHI) · PCI data · Data classification