AWS Certified AI Practitioner (AIF-C01) — Question 98

Which technique can a company use to lower bias and toxicity in generative AI applications during the post-processing ML lifecycle?

Answer options

• A. Human-in-the-loop
• B. Data augmentation
• C. Feature engineering
• D. Adversarial training

Correct answer: A

Explanation

The correct choice, Human-in-the-loop, involves human oversight and intervention in the machine learning process, which helps in identifying and mitigating biases and toxic outputs. The other options, while useful in different contexts, do not specifically address bias reduction in the post-processing phase of the ML lifecycle.