The author's argument would be weakened most by evidence that
Supporters of algorithmic sentencing tools point out that human judges are inconsistent: two defendants with similar histories can receive very different sentences depending on the judge, the time of day or the weather. A statistical tool, by contrast, treats like cases alike. The appeal of this consistency is real, but it conceals a trap. A model trained on past decisions inherits their biases, and then applies them uniformly and invisibly. A biased judge can be challenged in open court; a biased formula, protected as a trade secret, cannot. Consistency is a virtue only when what is being repeated is sound. Unless the workings of such tools can be inspected by the people they judge, we shall have exchanged variable injustice for standardised injustice.
Answer and explanation
Correct answer: A. the tools' predictions are demonstrably free of the biases found in past sentencing decisions
- Step 1: The author's warning depends on the tool inheriting bias from past decisions.
- Step 2: If the tools were demonstrably free of that bias, the warning about repeated injustice would lose force.
Answer
the tools' predictions are demonstrably free of the biases found in past sentencing decisions.
Why not the others: Judges' moods, the value of trade secrets, defendants' reading habits and the range of past sentences do not bear on whether the tool inherits bias, which is what the warning depends on.
Practise more Technology and justice.
Free questions with an explanation for every one, and full timed LNAT mock tests.