IAPP Artificial Intelligence Governance Professional (AIGP) — Question 145

A company developing and deploying its own AI model would perform all of the following steps to monitor and evaluate the model’s performance EXCEPT:

Answer options

• A. Publicly disclosing data with forecasts of secondary and downstream harms to stakeholders.
• B. Setting up automated tools to regularly track the model’s accuracy, precision and recall rates in real-time.
• C. Implementing a formal incident response plan to address incidents that may occur during system operation.
• D. Establishing a regular schedule for human evaluation of the model’s performance, including qualitative assessments.

Correct answer: A

Explanation

The correct answer is A because publicly disclosing data with potential harms is not a standard practice for evaluating AI model performance. Options B, C, and D are all legitimate steps in monitoring model performance, focusing on accuracy tracking, incident management, and human evaluation.