Study Shows AI Biases Can Unconsciously Sway Human Judgment

AI programs can display biases from their training data, which users may unconsciously absorb and apply in their own decision-making.

A study found people incorporated biases from a fake "AI" system into future judgments, even after no longer using it.

Biased AI could have serious impacts, like wrongly arresting more people of color with police facial recognition.

People may give more credibility to AI, allowing machine biases to more strongly sway human judgments.

Lack of transparency around AI training data makes it hard to detect inherent biases, so users should be educated on limitations.