Set 2: Inferences (Advanced)
Explanation
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?
Detailed Explanation
This question asks you to draw a logical conclusion from the text. Not 'programmed to discriminate' but learns from biased data = neutral-seeming systems embed bias. A valid inference must be supported by evidence in the passage, even if not stated directly. Look for clues in the text that strongly suggest the answer. Avoid conclusions that require assumptions beyond what's written. Valid inferences are strongly supported by multiple pieces of evidence in the text. Be cautious of choices that go too far beyond what the passage actually states. The best inference is the one most directly supported by textual evidence.
Key Evidence:
• "perpetuate and amplify existing biases"
• "data that already reflect discrimination"
Why others are wrong: A (Data 'already reflect discrimination.'), B (They 'perpetuate biases.'), C (They amplify, not correct, discrimination.).