Skip to content

CS Professional · Artificial Intelligence, Data Analytics and Cyber Security - Laws and Practice · Artificial Intelligence - Introduction and Basics

A Hyderabad company deploys a hiring-screening model trained on ten years of past hiring data in which most selected candidates were men. The model begins ranking women lower even though gender is not an input feature. Which statement best explains this and the appropriate lifecycle response?

Excluding gender does not remove bias, since proxy variables correlated with gender can carry historical bias from past hiring data. The sound response is to review the data and test outcomes across groups before deployment and keep monitoring afterwards, rather than assuming exclusion solves it.

  1. AProxy variables correlated with gender in the data can carry historical bias, so bias testing and data review should occur before and after deploymentCorrect
  2. BThe model cannot be biased because gender was excluded, so no further testing is needed
  3. CThe bias arises only from hardware faults and is solved by upgrading servers
  4. DBias is resolved automatically once the model is retrained on the same data

Explanation

Removing a protected attribute does not remove bias, because other features such as college, postcode or career gaps may act as proxies and encode past patterns. Responsible lifecycle practice therefore includes examining training data, testing outcomes across groups and monitoring after deployment. Retraining on the same biased data would only reproduce the problem.

Did you get it right without looking?

One question tells you little. A timed set on Artificial Intelligence - Introduction and Basics shows your real accuracy, how long you take and where you lose marks.

Practise Artificial Intelligence - Introduction and BasicsFree diagnostic mock

More Artificial Intelligence - Introduction and Basics questions