Model or data drift: Which of the following is a recognized technical cause they should examine?

Model or data drift, where production inputs diverge from the training distribution, is a recognized technical cause of degraded predictions.

The question

Preparing for AIGP? Take the free 5-min readiness quiz →

1. Model or data drift, where changes in real-world input distributions gradually diverge from the data the model was trained on. ✓

   Correct: drift is a recognized cause of degraded performance, as production inputs move away from the training distribution.
2. The absence of a formal press release, which may slow the system's adoption but does not change how it performs on real inputs.

   Plausible as a project gap, but communications do not affect prediction quality.
3. A shortage of graphical dashboards, which can limit the team's visibility but is not itself a cause of degraded prediction quality.

   Plausible since dashboards aid monitoring, but their absence does not degrade the model's outputs.
4. The vendor's choice of programming language, which mainly affects developer productivity rather than the correctness of predictions.

   Plausible as an engineering factor, but language choice does not by itself cause poor predictions.

How to remember it

Model or data drift, where production inputs diverge from the training distribution, is a recognized technical cause of degraded predictions.

More Understanding How to Govern AI Development questions

• Record the incident and act to contain it: Following good post-deployment governance, what should the team do →
• Technical documentation and instructions for use that tell: Which disclosure most directly serves that →
• It fixes the problem: Why does defining the business context matter before feature engineering? →
• All 426 Understanding How to Govern AI Development questions →