Country
You are in the Nigeria edition. Type your country to switch.
Nigeria
Loading every country…
No country matches. Check the spelling, or choose "Any country" for the Global edition.
Your learning profile
AI safety and security at work · Lesson 3 of 9
Data leakage
Data leaks through AI tools by prompts, uploads, connectors, outputs and logs, usually during ordinary work. Classification labels, approved tools and data loss prevention reduce the risk, and a worked month of incidents shows where to act first.
This lesson is complete in writing. A video for it is on its way.
What you will learn
- Name the five routes by which data leaks through AI tools: prompts, uploads, connectors, outputs and logs.
- Explain shadow AI and why approved tools matter.
- Use classification labels to decide what data may go into which tool.
- Describe what data loss prevention can and cannot catch, and read a month of incident figures.
Key terms in this lesson
- Sensitive data exposure
- Sensitive information escaping where it should not go, for example through prompts, uploads, connected tools, logs or AI outputs.
- Shadow AI
- Staff using AI tools that the organisation has not approved, often with company, customer or candidate data, outside its security, privacy and quality controls.
- Data classification
- Sorting information into a few classes, such as public, internal, confidential and restricted, so the strongest protection goes to the data whose leak would cause most harm.
- Data loss prevention
- Tools and rules that detect sensitive data by pattern, label or destination and warn about, block or log it before it leaves an organisation.
- Audit trail
- A record that lets someone trace a published figure back through every calculation to its source, and see when and why it changed.
- Step 1: Five routes for data to leak, not started
- Step 2: Shadow AI and approved tools, not started
- Step 3: Classification labels decide what goes where, not started
- Step 4: Data loss prevention: what it catches and misses, not started
- Step 5: Worked review: a month of incidents, not started
About 10 min of reading
Step 1 of 5
Five routes for data to leak
Opening this step…
Step 2 of 5
Shadow AI and approved tools
This step opens when you finish the one above.
Step 3 of 5
Classification labels decide what goes where
This step opens when you finish the one above.
Step 4 of 5
Data loss prevention: what it catches and misses
This step opens when you finish the one above.
Step 5 of 5
Worked review: a month of incidents
This step opens when you finish the one above.
Key takeaways
The key takeaways open when you finish every step.
Lesson quiz
5 questions. You need 85% to pass.
The quiz opens when you finish every step.
Read first
Stockertim Learn is education. It explains how investing works and never recommends buying or selling anything.
Popular this week: Basics: BeginnerStock market: Intermediate