The model may confidently propose unsupported routes: Which residual risk still most specifically requires
Fresh retrieval does not guarantee faithful output; the model may still generate unsupported route recommendations.
The question
A logistics company proposes replacing manual route scheduling with a generative AI system. It maps users, data access, dependencies, and controls. A pilot adds a timestamped traffic feed with freshness monitoring and retrieval. Which residual risk still most specifically requires attention?
Preparing for AIGP? Take the free 5-min readiness quiz →
- The system may need additional cloud capacity.Capacity planning is a legitimate deployment concern, but it does not address whether generated schedules are supported by the available evidence.
- The model may confidently propose unsupported routes despite retrieval. ✓Retrieval and freshness monitoring improve available context but do not eliminate confabulation or guarantee faithful route recommendations.
- Dispatchers may need authority to approve route changes.Approval authority is an important operational control, but it is distinct from the residual risk that generated routes may be unsupported.
- The traffic feed may require a different vendor contract.Supplier terms may require review, but the stated freshness controls do not guarantee that generated route recommendations are faithful to the retrieved data.
The trap
Ask what a mitigation actually controls; retrieval does not by itself guarantee accurate or supported output. How to remember it
Fresh retrieval does not guarantee faithful output; the model may still generate unsupported route recommendations.
How many of these would you get right?
One of 1581 AIGP questions on Certsqill. Take a free five-minute check and see your score per domain — not one number, but which section to open tonight.
Test your AIGP readiness — freeMore Understanding How to Govern AI Deployment and Use questions
- Enforce approval before the tool can submit: The team proposes relying on a system instruction saying, “Do not →
- Reviewers cannot intervene effectively: Which remaining workforce-readiness risk is most decisive? →
- Local validation is still required: What evidence gap remains decisive before deployment? →
- All 424 Understanding How to Govern AI Deployment and Use questions →
Part of the Certsqill AIGP question bank · Understanding How to Govern AI Deployment and Use ·
Every answer, right and wrong, comes with its own explanation.