5

Set 2: Inferences (Advanced)

Explanation

Answer: B

PASSAGE

Machine learning systems trained on historical data can perpetuate and even amplify existing biases. For example, a hiring algorithm trained on past decisions may learn to discriminate against groups that were previously underrepresented. The system isn't programmed to discriminate—it infers patterns from data that already reflect discrimination.

What does the passage suggest about algorithmic decision-making?

A. Algorithms are inherently neutral and unbiased.
B. Apparently neutral technical systems can embed and perpetuate societal biases through data.✓ Correct
C. Machine learning corrects discrimination automatically.
D. Historical data never contains biases.

Detailed Explanation

Not 'programmed to discriminate' but learns from biased data = neutral-seeming systems embed bias.

Key Evidence:

• "perpetuate and amplify existing biases"

• "data that already reflect discrimination"

Why others are wrong: A (They 'perpetuate biases.'), C (They amplify, not correct, discrimination.), D (Data 'already reflect discrimination.').

← PreviousNext →