NCA-GENL Trustworthy AI Practice Question
A media company fine-tunes an NVIDIA Nemotron model on licensed news articles to build a summarization tool. Legal asks how the team can demonstrate that the training data was lawfully acquired and that the model does not reproduce copyrighted passages verbatim. Which combination of practices best addresses both concerns?
⚠ Common exam trap
The trap here is treating differential privacy or a user-facing disclaimer as a complete answer, when neither documents data origin nor measures verbatim reproduction.
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
✓
Maintain a data provenance record for each training source and run memorization tests that check for verbatim reproduction of training passages.
Data provenance records and memorization testing together provide auditable evidence for both legal concerns. Provenance documents origin and licensing, while memorization tests measure whether the model reproduces training passages verbatim. Neither alone is sufficient: provenance without testing leaves reproduction risk unmeasured, and testing without provenance leaves acquisition unverified.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Apply differential privacy during fine-tuning and assume it eliminates all copyright risk without further testing.
Why it's wrong here
Differential privacy can reduce the risk of memorization, but it does not document where training data came from, so it cannot answer the lawful-acquisition question. Assuming it eliminates all copyright risk without memorization testing removes the evidence legal needs. It is a partial mitigation at best and does not address provenance.
- ✗
Add a disclaimer to the tool's user interface stating that summaries may resemble source articles.
Why it's wrong here
A disclaimer does not demonstrate lawful data acquisition and does not measure whether the model reproduces copyrighted text. Legal asked for evidence, not a warning label. This option shifts risk to the user without providing the provenance records or memorization testing that would actually satisfy the legal review.
- ✓
Maintain a data provenance record for each training source and run memorization tests that check for verbatim reproduction of training passages.
Why this is correct
Data provenance records document the origin, license, and chain of custody for each training source, directly answering the lawful-acquisition question. Memorization tests probe whether the model can reproduce training passages verbatim, providing evidence about copyright risk. Together they address both legal concerns with concrete, auditable artifacts that the legal team can review.
- ✗
Increase the model's parameter count and retrain on a larger unlicensed web crawl to improve summarization quality.
Why it's wrong here
Scaling parameters and using an unlicensed web crawl worsens the legal exposure rather than addressing it. The scenario specifically requires demonstrating lawful acquisition and low verbatim reproduction, and an unlicensed crawl provides neither. This choice prioritizes quality metrics over the compliance requirements the legal team raised.
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.