What is a key practice for ensuring fairness in AI?

• A. Implement bias mitigation strategies and monitor outcomes to ensure fairness.
• B. Increase data collection without regard to quality.
• C. Increase model complexity without testing.
• D. Ignore post-deployment monitoring.

Answer: (A)

Ensuring fairness in AI requires actively mitigating bias and continuously monitoring outcomes. Implementing bias mitigation strategies helps reduce discriminatory patterns in data and model behavior, while ongoing monitoring across different groups reveals disparities, drift, or unexpected impacts in production. This combination creates a proactive, verifiable approach to fairness by enabling timely adjustments and governance checks, rather than relying on one-time fixes.

Why the other options don’t fit: collecting more data without regard to quality can amplify biases if the data isn’t representative; increasing model complexity without proper testing can reduce transparency and governance, making fairness harder to assess; and ignoring post-deployment monitoring leaves you blind to real-world issues that emerge after deployment.