Which type of bias involves discrimination toward a specific group due to skewed data?

The identified answer relates to implicit bias, which is fundamentally about the attitudes or stereotypes that affect our understanding, actions, and decisions in an unconscious manner. Implicit bias occurs when we make judgments about individuals based on their membership in a specific group, which can be influenced by societal norms, stereotypes, and personal experiences.

In the context of AI and data-driven systems, this type of bias can manifest when the training data reflects historic discrimination or unequal representation of certain groups. If a dataset includes disproportionate examples of a particular demographic, the AI systems trained on that data may inadvertently perpetuate those biases, leading to discriminatory outcomes based on skewed data.

Temporal bias refers to issues that arise due to the time-related changes in data, sampling bias pertains to the way samples are collected, affecting representativeness, and overfitting relates to a model's performance on training data versus new data. While these concepts are important to understand in the field of AI governance, they do not specifically address the discrimination toward a particular group due to biased data in the same direct manner that implicit bias does. Therefore, implicit bias is the most accurate choice regarding discrimination stemming from skewed data.