A project manager is tasked with overseeing the implementation of an AI model for financial forecasting. They need to ensure the model's predictions are reliable. If the model's error rate exceeds acceptable boundaries, what will occur next?

In the AI project lifecycle, a model must pass a rigorous evaluation phase before it can be deployed or "operationalized." If the model's performance, measured by metrics like the error rate, does not meet the pre-defined acceptance criteria, it is considered unfit for production. The standard corrective action is to return to the model development and training phase. This iterative process of refining data, tuning hyperparameters, and retraining the model to improve its accuracy inevitably consumes time, leading to a direct delay in the project's deployment schedule.

B. A high error rate indicates unreliability, which would necessitate increased human oversight and validation, not a reduction.
C. While retraining does incur computational costs, the most immediate and significant project management impact is the schedule delay in operationalization.
D. Discovering a high error rate would likely decrease stakeholder confidence in the project's current state, not increase it.

1. Amershi, S., Begel, A., Bird, C., DeLine, R., Gall, H., Kamar, E., Nagappan, N., Nushi, B., & Zimmermann, T. (2019). Software Engineering for Machine Learning: A Case Study. 2019 IEEE/ACM 41st International Conference on Software Engineering: Software Engineering in Practice (ICSE-SEIP), 291-300. DOI: 10.1109/ICSE-SEIP.2019.00042. (This paper describes the nine stages of an ML workflow, highlighting its iterative nature. On page 295, it shows that after "Model Evaluation," if the model is not satisfactory, the process loops back to "Model Training" or earlier stages, inherently causing delays before "Model Deployment.")

2. Lwakatare, K., Raj, A., Crnkovic, I., Bosch, J., & Olsson, H. H. (2020). Large-scale machine learning systems in practice: a case study of challenges and lessons learned. Information and Software Technology, 127, 106378. DOI: 10.1016/j.infsof.2020.106378. (Section 4.2, "Model development and experimentation," discusses the challenge of models not meeting performance requirements, which leads to further experimentation and iteration, thus delaying deployment.)

3. Project Management Institute. (2024). The Project Manager's Guide to Mastering AI. PMI White Paper. (This guide emphasizes the unique, iterative, and experimental nature of AI projects. It implicitly supports the concept that unmet performance metrics require revisiting development phases, which impacts project timelines, a core concern for project managers.)