QuestionQ35

AI Risk Management

An attacker gradually alters portions of the training data in an organization’s developing AI system. If the manipulation is not detected, which of the following is the MOST likely risk?

Explanation

Training-data manipulation can poison the patterns a model learns, leading it to generate distorted, unfair, or systematically biased decisions.

Community Discussion

No comments yet. Be the first to start the discussion!