Which term describes the presence of unfair or prejudiced outcomes in AI systems due to biases in data, algorithms, or design?

• A. Bias
• B. Ethics
• C. Generated Knowledge Prompting
• D. Iterative Prompting

Answer: (A)

Bias in AI refers to the presence of unfair or prejudiced outcomes that arise when data, algorithms, or design choices reflect or amplify societal prejudices. If training data underrepresents a group, the model may perform worse for that group, leading to discriminatory predictions. Similarly, optimization that focuses on overall accuracy can hide harms to minority groups, producing biased results. Design decisions, such as which features to include or how thresholds are set, can also embed these biases into the system. This is why the term that describes this phenomenon is bias. Ethics deals with what ought to be, not the description of unfair outcomes, and the other two options relate to prompting methods rather than fairness or bias.