Match the AI Ethics with its correct description. AI Ethics Description I Data…
2024
Match the AI Ethics with its correct description.
AI Ethics | Description | ||
I | Data privacy | 1 | It is a standard metric that measures how many users in a database share a particular attribute. |
II | K-anonymity | 2 | It is a trade-off between the amount of data collected and the performance of a product or service. |
III | L-diversity | 3 | It extends the idea by merging several sensitive attributes together. |
- A.
I-1, II-3, III-2
- B.
I-2, II-3, III-1
- C.
I-1, II-2, III-3
- D.
I-2, II-1, III-3
Attempted by 96 students.
Show answer & explanation
Correct answer: D
Concept
This question pairs three privacy ideas with their descriptions. In the abstract: data privacy is the overall practice of balancing how much data is collected against the usefulness or performance you obtain from it. K-anonymity is a group-based guarantee — records are grouped so that each group holds at least k records, making any single record indistinguishable from at least k-1 others; it is therefore characterised by how many users share the same attribute pattern. L-diversity is built on top of k-anonymity, adding the requirement that the sensitive attributes within each such group be diverse.
Applying it to each pair
Data privacy is the description about a trade-off between the amount of data collected and product or service performance. This collect-less-versus-perform-better tension defines data privacy as a whole, not any single anonymisation model.
K-anonymity is the description about a metric counting how many users in a database share a particular attribute. The parameter k is precisely that group size: each record sits in a group of at least k records sharing the same pattern.
L-diversity is the description that extends the idea over sensitive attributes. It is layered on k-anonymity so that a group does not leak a sensitive value when all of its members happen to share it.
Resulting matches
AI Ethics | Matched description |
|---|---|
Data privacy | 2 — data-collection vs. performance trade-off |
K-anonymity | 1 — how many users share an attribute |
L-diversity | 3 — extends the idea over sensitive attributes |
Cross-check and a note on wording
Each idea maps to exactly one description with none left over, giving I-2, II-1, III-3. Note that description 3 phrases l-diversity loosely as 'merging several sensitive attributes together'; strictly it requires diversity of sensitive values within a group, but among the three descriptions it is still the only one that captures l-diversity as an extension over sensitive attributes.