AIF-C01 Fundamentals of Generative AI Practice Question

A media company is evaluating foundation models for an application that will summarize long earnings-call transcripts in English for internal analysts. The transcripts are up to 40,000 words, and the summaries must remain faithful to the source. Which TWO model characteristics are most important to evaluate for this workload? (Choose two.)

⚠ Common exam trap

The trap here is selecting impressive-sounding multimodal or edge capabilities instead of the two characteristics that actually govern long-document summarization quality.

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

✓

The model's quality on text summarization and faithfulness to source content

Long transcripts demand a model whose context window can accommodate the input, and financial summarization demands strong faithfulness so the output does not distort the source. Those two characteristics together determine whether the application can produce trustworthy summaries without excessive chunking. Image generation, speech quality, and edge deployment do not address the stated requirements.

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

✗
The model's ability to run on edge devices without network connectivity
Why it's wrong here
The scenario describes an internal analyst tool, and nothing indicates offline or edge deployment requirements. Large-context summarization models generally require substantial compute and are served through managed APIs rather than on edge hardware. Edge capability is therefore not a relevant evaluation axis here.
✗
The model's text-to-speech voice quality
Why it's wrong here
Voice quality matters only if the summaries are read aloud to users, and the scenario specifies internal analysts reading summaries. Text-to-speech is a separate capability typically provided by services such as Amazon Polly rather than a foundation model selection criterion. It does not affect summarization accuracy or coverage.
✓
The model's quality on text summarization and faithfulness to source content
Why this is correct
Faithfulness determines whether the summary reflects what the transcript actually said rather than inventing figures or positions. A model with weak summarization quality will produce fluent but unreliable output, which is unacceptable for financial analysis. Benchmarking summarization quality on representative transcripts is therefore essential.
✗
The model's ability to generate images from text
Why it's wrong here
The requirement is textual summarization of transcripts, so image generation capability is irrelevant and would not improve summary fidelity. Including a multimodal generation feature adds cost and complexity without addressing the task. Evaluators should focus on text quality and input capacity instead.
✓
The model's supported context window size
Why this is correct
With transcripts up to 40,000 words, the model must be able to accept a very large input in a single request, or the application must chunk the text. Context window size directly determines whether the full transcript fits, and it constrains how much surrounding context the model can use. It is a primary selection criterion for long-document summarization.

Go deeper

AIF-C01 Fundamentals of Generative AI Practice Question

Why each option matters

Option-by-option breakdown

Related to this question