AAISM: Advanced in AI Security Management

By Isaca · Question Mode

QuestionQ87

AI Risk Management

After deployment, an AI model’s output starts drifting beyond its expected range. What is the development team’s BEST course of action?

Explanation

Model monitoring is an iterative part of the AI life cycle. Output drift should trigger a return to the appropriate earlier phase to investigate changes in data or operating conditions, reassess the model, and update or retrain it as needed.

Community Discussion

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