Is AI Racist? Sometimes. | AI 103: Ethics (Part 1 of Many)

What is the primary reason AI can produce discriminatory outcomes despite lacking intent?

• AI systems have a consciousness that influences decisions.

• AI systems are programmed to discriminate.

• AI systems are inherently biased.

• AI systems rely on data that may contain biases.

In the recidivism prediction example, what factor indirectly led to racial bias in the software's risk assessment?

• Criminal justice system records

• Socioeconomic history

• Defendant's age

• Direct questions about race

What is the main challenge when AI systems use correlations in data?

• Correlations are irrelevant to AI.

• Correlations are easy to identify.

• Correlations can be mistaken for causation.

• Correlations are always accurate.

How can AI biases in facial recognition be addressed?

• By using more complex algorithms

• By increasing the number of datasets

• By using representative datasets

• By ignoring the biases

What is a recommended approach to ensure AI predictions are fair across different groups?

• Use a single dataset for all predictions

• Avoid using AI for sensitive tasks

• Validate predictions against trends

• Rely solely on historical data