5.6 - Evaluating AI Outputs

Question 1

Why is evaluating AI outputs important?

Answer 1

AI outputs can be inaccurate, biased, or misaligned with the problem if not assessed carefully.

Question 2

What does evaluation of AI outputs involve?

Answer 2

Checking outputs for accuracy, consistency, relevance, bias, and ethical considerations.

Question 3

Name two dimensions to assess in AI outputs.

Answer 3

Examples: Accuracy, relevance, consistency, fairness, interpretability.

Question 4

Comparing AI predictions against actual outcomes to measure performance is called ______.

Answer 4

Validation / testing

Question 5

Give one example of a metric to evaluate predictive AI models.

Answer 5

Examples: RMSE (Root Mean Square Error), accuracy, precision, recall.

Question 6

What is one way to check for bias in AI outputs?

Answer 6

Example: Compare predictions across different groups or demographics.

Question 7

Which of these is NOT a best practice for evaluating AI outputs?
A) Cross-checking with known data
B) Ignoring inconsistent results
C) Documenting assumptions
D) Monitoring over time

Answer 7

B) Ignoring inconsistent results

Question 8

Why is interpretability important when evaluating AI outputs?

Answer 8

So stakeholders can understand why a model produced a certain result and trust the recommendations.

Question 9

Continuous monitoring of AI outputs to detect drift or errors is called ______.

Answer 9

Model monitoring / post-deployment evaluation

Question 10

Name one tool or technique to evaluate AI outputs.

Answer 10

Examples: Confusion matrix, error analysis, dashboards for tracking metrics.

Question 11

How can evaluation help improve AI models?

Answer 11

By identifying weaknesses, biases, or errors, allowing for refinement and better performance.

Question 12

How does evaluating AI outputs relate to responsible AI?

Answer 12

Ensures AI decisions are fair, accurate, transparent, and aligned with ethical standards.