A large language model deployed in a legal research application generates a case citation that appears authoritative but does not exist. A lawyer relies on the citation in court documents. Which risk category does this PRIMARILY represent?
Select an answer to reveal the explanation.
Short Explanation and Infographic
Here's the deal — b is correct because LLM hallucination — generating confident but factually incorrect outputs, including fabricated citations — is a well-documented generative AI risk that creates significant liability risk when deployed in professional domains like legal research. A requires evidence of interception.
Full explanation below image
Full Explanation
B is correct because LLM hallucination — generating confident but factually incorrect outputs, including fabricated citations — is a well-documented generative AI risk that creates significant liability risk when deployed in professional domains like legal research. A requires evidence of interception. C requires evidence that fabricated data was in training. D requires evidence of adversarial prompting.