AWS Certified AI Practitioner · Free Practice Question Medium
Question 40
A media company uses a generative AI model to auto-generate news summaries. They notice that some outputs include fabricated quotes attributed to public figures, which closely resemble their speaking style but contain factual inaccuracies. Which challenge of generative AI does this scenario BEST illustrate?
-
A
Hallucinations producing plausible but incorrect content
-
B
High latency during model inference
-
C
Limited scalability for large documents
-
D
Bias in training datasets affecting fairness
Reveal correct answer
Correct answer: A
Explanation
Hallucinations Producing Plausible but Incorrect Content
Generative AI models can create text that looks coherent and believable, but is not based on factual information. When models generate details—such as invented quotes or false data—this occurs because they rely on patterns found in their training data, rather than a grounded understanding of real-world facts. These “hallucinations” pose a significant risk in areas like journalism, where factual accuracy is critical.
Why Other Challenges Are Less Relevant in This Scenario
High Latency During Model Inference
This challenge would manifest as noticeable delays in generating responses. The problem here is not speed; it is the fabrication of incorrect quotes.
Bias in Training Datasets Affecting Fairness
Bias arises when models systematically produce skewed or discriminatory outputs. Although bias can lead to unfair or offensive content, it does not directly explain the presence of fabricated quotes attributed to public figures.
Limited Scalability for Large Documents
This challenge pertains to difficulties handling or generating content for very large volumes of text. In the described scenario, there’s no indication that the model struggles with scaling; the issue is the production of inaccurate details.
In summary, the media company’s situation—fabricated yet seemingly believable quotes—best exemplifies the generative AI challenge of producing coherent but factually incorrect content, commonly termed “hallucinations.”
A. This scenario best illustrates the challenge of hallucinations producing plausible but incorrect content in generative AI models. The fabricated quotes attributed to public figures closely resembling their speaking style but containing factual inaccuracies demonstrate how the model can generate content that is believable but ultimately false.
B. High latency during model inference refers to delays in processing and generating outputs from the AI model. While latency issues can impact the user experience, they are not directly related to the scenario of fabricated quotes with factual inaccuracies.
C. Limited scalability for large documents refers to the model's ability to handle and process large amounts of data efficiently. While scalability is an important consideration in AI systems, it is not the main challenge highlighted in the scenario of fabricated quotes with factual inaccuracies.
D. Bias in training datasets affecting fairness pertains to the presence of biases in the data used to train the AI model, which can result in unfair or discriminatory outcomes. While bias is a significant challenge in AI, it is not the primary issue in the scenario of fabricated quotes with factual inaccuracies.
Discussion
Think the marked answer is wrong, or have a better explanation? Share it below — comments appear after review.
