Exam AB-730 Topic 1 Question 35 Discussion

In Microsoft 365 Copilot, when a generated report includes fictional or factually incorrect information that sounds plausible, it is primarily called a hallucination. While some technical discussions use terms like fabrication to describe the act of inventing details, "hallucination" is the industry-standard term for this phenomenon.
Key characteristics of these errors include:
Hallucination: Occurs when the AI generates content that is not grounded in the provided data or real-world facts, often appearing coherent but being entirely inaccurate.
*-> Fabricated Answers: The model may "fill in the blanks" with invented numbers, dates, or names when it lacks specific information from its sources.
Ungrounded Content: Microsoft specifically uses this term to describe information that does not align with the connected data sources, such as web info or internal documents.
Confabulation: An alternative metaphor sometimes used to describe the creative "gap-filling" process the AI performs when it lacks data.
Incorrect:
[Not A]
Deepfakes specifically refer to synthetic media-such as images, videos, or audio recordings- that have been manipulated to mimic a real person's likeness or voice. These are typically created deliberately to deceive or mislead.
[Not E]
Bias refers to the tendency of the AI to generate outputs that unintentionally reflect, reinforce, or amplify harmful stereotypes, societal inequalities, or prejudices present in its training data.
Because Copilot is a generative AI tool trained on vast amounts of internet text and user data, it can inadvertently produce content that is skewed toward specific viewpoints, cultures, or demographics.
Reference:
https://www.crmsoftwareblog.com/2026/02/speak-ai-fluently-a-complete-guide-to-understanding- microsoft-365-copilot-terminology/