Personal data
Personal data is information relating to an identified or identifiable person, directly or indirectly, alone or in combination with other information.
Introduction to Generative AI, Second Edition
Personal data is information relating to an identified or identifiable individual, including information that can identify the person directly or indirectly.
- How it is used
- Frames privacy risks in both LLM training data and users’ conversations with AI services.
- How deep it goes
- Explains identification, memorization, practical safeguards, and legal obligations at an overview level without prescribing an implementation.
- Examples
- Uses extracted GPT-2 training sequences, chatbot conversations, and student assignments to make exposure risks concrete.
- What you can do after
- Evaluate/understand where personal data may enter an AI system, how it might be exposed, and why safeguards need scrutiny.
Read how Introduction to Generative AI, Second Edition explains personal data →
Effective Data Analysis
Personal data, also called personal information or personally identifiable information, is information that identifies or can reasonably be linked to an individual.
- How it is used
- Guides privacy-conscious decisions about collecting, analyzing, sharing, and retaining information that identifies individuals.
- How deep it goes
- Defines identifying data and works through field removal, geographic reduction, and pseudonymization without prescribing one compliance procedure for every jurisdiction.
- Examples
- Uses website metadata and a transaction table containing names, phone numbers, addresses, and emails.
- What you can do after
- Apply it via pandas to remove unnecessary identifiers, retain less granular location data, and separate hashed email identifiers from originals.
Data for All
Personal data is information relating to an identified or identifiable individual.
- How it is used
- Frames choices about collecting, sharing, protecting, and potentially licensing information about individuals.
- How deep it goes
- Explains practical boundaries through Open Data and deidentification, without providing a formal classification test for personal data.
- Examples
- Uses wearable health measurements, social media activity, account records, and a DNA deidentification dispute.
- What you can do after
- Evaluate/understand how everyday records reach organizations and why sharing or removing obvious identifiers calls for care.
Machine Learning Platform Engineering
Personally identifiable information is information that can identify or be used to identify an individual.
- How it is used
- Protects LLM pipelines by detecting and sanitizing identifying information in user inputs.
- How deep it goes
- Defines identifying information and demonstrates validator configuration, redaction, exception handling, and sanitized output.
- Examples
- Uses email addresses, credit-card numbers, Social Security numbers, and automatically redacted prompt values.
- What you can do after
- Apply it via Guardrails AI to validate, redact, reject, and log sensitive user input before LLM processing.
Read how Machine Learning Platform Engineering explains personal data →
The Art of AI Product Development
Personal information is information that relates to or can identify an individual.
- How it is used
- Guides AI privacy, governance, compliance, stakeholder communication, and training-data evaluation.
- How deep it goes
- Defines personal data, explains privacy controls and risks, and stops short of detailed implementation procedures.
- Examples
- Uses a ChatGPT data-leakage incident, anonymized customer orders, regulatory rights, and scraped language-model training data.
- What you can do after
- Evaluate AI data practices by checking consent, minimization, isolation, access controls, retention, provenance, and transparency.
Read how The Art of AI Product Development explains personal data →
First-Party Data Activation
Data that identifies, relates to, describes, or can reasonably be linked to an individual.
- How it is used
- Supports personalization, marketing analysis, fulfillment, experience optimization, and demand forecasting while requiring privacy governance.
- How deep it goes
- Defines reasonable linkage and addresses consent, re-identification, and controls from intake through processing.
- Examples
- Uses profile attributes, interaction records, transactional data, sessions, and fitness recommendations as concrete examples.
- What you can do after
- Operate privacy and security controls at intake and throughout subsequent processing.
Read how First-Party Data Activation explains personal data →
Starting Data Analytics with Generative AI and Python
Personally identifiable information is information that can identify a specific individual directly or when combined with other information.
- How it is used
- Guides decisions about whether data may be uploaded to generative AI platforms for analysis.
- How deep it goes
- Defines identifying information and gives practical screening, anonymization, oversight, and accountability guidance.
- Examples
- Uses uploaded text, CSV, or Excel files and a warning about apparently non-sensitive proprietary algorithms.
- What you can do after
- Evaluate whether data requires anonymization, approval, restricted handling, or exclusion before AI-assisted analysis.
Read how Starting Data Analytics with Generative AI and Python explains personal data →
Grokking AI Algorithms, Second Edition
Personally identifiable information is data that can identify or be linked to a specific individual, such as a phone or credit-card number.
- How it is used
- Removes identifying information during language-model training-data cleaning to address safety, privacy, and compliance risks.
- How deep it goes
- Defines PII and explains why to strip it, but stops short of detection algorithms or implementation details.
- Examples
- Uses phone numbers, credit-card numbers, email addresses, and ID numbers as examples of sensitive data.
- What you can do after
- Apply data preprocessing that removes personal data before training a language model.
Read how Grokking AI Algorithms, Second Edition explains personal data →
Building Reliable AI Systems
Information relating to an identified or identifiable person.
- How it is used
- Protects information relating to identifiable people through a proactive privacy layer in responsible AI systems.
- How deep it goes
- Defines personal data and explains exposure risks and compliance stakes, but does not specify detection or de-identification procedures.
- Examples
- Contrasts a Samsung ChatGPT data leak with HIPAA identifier types and GDPR consent and erasure requirements.
- What you can do after
- Understand why privacy protection requires proactive safeguards, compliance measures, and ongoing production monitoring.
Read how Building Reliable AI Systems explains personal data →
Related concepts
- sensitive informationkind of · Sensitive information is information requiring protection because its unauthorized access, disclosure, alteration, or misuse could cause harm or violate obligations.
- GDPRconfigured by · The General Data Protection Regulation (GDPR) is a European Union regulation governing the processing of personal data and protecting individuals’ rights concerning that data.
- open datasetcontrasts with · An open dataset is a dataset made publicly accessible for use and reuse under its stated legal and technical conditions.
- Facebookused by · Facebook is an online social networking platform for creating profiles, connecting with users, publishing and sharing content, and interacting with audiences.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.