Attachments
  • Mark as Completed
  • View
    17
  • More
Previous class

Section 2: No personal data

Module 7 — Data Safety & Common Mistakes

Section 2: No Personal Data

Purpose of This Section

This section explains why personal data should not be entered into AI tools and how even small details can create serious privacy and compliance risks.

Personal data exposure is cumulative, not isolated.

The Core Idea

Personal data connects dots.

Even individual details that seem harmless—names, emails, job titles, dates—can become identifying when combined with context. Once personal information is shared, control over how it is used or inferred is reduced.

AI does not need personal data to be effective.

What Counts as Personal Data

Personal data includes any information that can identify or reasonably be linked to a real person, such as:

  • names or initials
  • email addresses or phone numbers
  • employee or customer IDs
  • resumes or performance reviews
  • complaints or feedback tied to individuals
  • combinations of role, location, and timing

Even partially anonymized data can become identifiable with enough context.

Why This Matters

Sharing personal data creates:

  • privacy violations
  • legal and regulatory risk
  • loss of trust
  • potential harm to individuals

These risks increase when data is shared casually or under time pressure.

What AI Is Safe to Use Instead

You can still use AI effectively by:

  • replacing names with placeholders
  • generalizing scenarios
  • removing identifying details
  • focusing on structure, tone, or intent
  • asking hypothetical or representative questions

AI works best with patterns, not identities.

Common Failure Mode

A common mistake is assuming that small details are safe.

Another failure mode is believing that internal tools or widely used platforms eliminate risk. Personal data exposure does not depend on intent—it depends on access and context.

Convenience does not reduce responsibility.

The Conjugo Rule

If it identifies a real person, do not share it.

When in doubt, remove or abstract the information before using AI.

Best Practices

Protecting personal data works best when:

  • prompts are written without identifiers
  • placeholders are used consistently
  • scenarios are generalized
  • sensitive material remains in controlled systems

Boundaries protect people and organizations.

Section Takeaway

  • Personal data creates compounding risk
  • Small details can become identifying
  • AI does not require personal data to help
  • Responsibility for privacy remains human

Protect personal data by default.

This concludes Section 2.