NCA-GENL Trustworthy AI Practice Question
A retail company's LLM-based product recommendation assistant begins suggesting discontinued items and outdated pricing roughly six weeks after launch, even though the model weights have not changed. The team confirms the training data and prompts are unchanged. Which phenomenon best explains the degraded output quality?
⚠ Common exam trap
The trap here is attributing gradual quality loss to a defect inside the model, when an unchanged model can only degrade because the world it was trained to represent has shifted.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Concept drift in the business environment
Because the model, prompts, and training data are static while recommendations degrade over weeks, the change must come from the environment the model describes. Discontinued items and updated prices alter the correct mapping from customer query to product, which is concept drift. The other choices require retraining, fixed preprocessing faults, or time-accumulating numeric errors, none of which fit an unchanged model that initially performed well.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Quantization error accumulation over time
Why it's wrong here
Quantization error is deterministic and fixed at conversion time; it does not grow week by week while the model sits idle. If precision loss were the cause, quality would be uniformly lower from deployment rather than deteriorating after six weeks. The observed timeline points to changes in the product catalog, not to numeric precision degrading on its own.
- ✓
Concept drift in the business environment
Why this is correct
Concept drift occurs when the statistical relationship between inputs and correct outputs changes over time because the real-world environment moved. Discontinued products and new prices mean the same customer query now maps to a different correct answer, so a frozen model trained on old catalog relationships becomes stale. This matches the six-week degradation with unchanged weights and prompts.
- ✗
Catastrophic forgetting during inference
Why it's wrong here
Catastrophic forgetting describes a model losing previously learned capabilities when fine-tuned on new data, which requires additional training. Here the weights are explicitly unchanged and no retraining occurred, so there is no mechanism for forgetting. The stale recommendations instead reflect a changing catalog, not a loss of learned knowledge inside the network.
- ✗
Tokenization mismatch in the embedding layer
Why it's wrong here
A tokenization mismatch would cause consistent failures from the first day, such as garbled handling of certain characters or terms, not a gradual six-week decline. Since the prompts and training data are unchanged and the system worked initially, the tokenizer is functioning as designed. The drift is temporal and environmental rather than a fixed preprocessing defect.
About these practice questions
Courseiva writes every NCA-GENL question from scratch — 367 in total, each with an explanation and a wrong-answer breakdown. None are copied from real exams or dumps. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official NVIDIA exam blueprint
This NCA-GENL practice question is part of Courseiva's free NVIDIA certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the NCA-GENL exam.