Hallucination
Despite the rapid advancement of LLMs, hallucinations have emerged as one of the most vital concerns surrounding their use [54, 79, 86, 110, 242]. Hallucinations are often referred to as LLMs’ generating content that is nonfactual or unfaithful to the provided information [54, 79, 86, 242]. Therefore, hallucinations can be typically categorized into two main classes. The first is factuality hallucination, which describes the discrepancy between LLMs’ generated content and real-world facts. For example, if LLMs mistakenly take Charles Lindbergh as the first person who walked on the moon, it is a factuality hallucination [79]. The second is faithfulness hallucination, which describes the discrepancy between the generated content and the context provided by the user’s instructions or input, as well as the internal coherence of the generated content itself. For example, when LLMs perform the summarizing task, they occasionally tamper with some key information by mistakes, which is a faithfulness hallucination.
ENTITY
2 - AI
INTENT
2 - Unintentional
TIMING
2 - Post-deployment
Risk ID
mit1512
Domain lineage
3. Misinformation
3.1 > False or misleading information
Mitigation strategy
1. Prioritize the implementation of Retrieval-Augmented Generation (RAG) to ground LLM outputs in verified, external knowledge sources, thereby enhancing factual accuracy and mitigating both factuality and faithfulness hallucinations. 2. Employ systematic Advanced Prompt Engineering techniques, such as using explicit instructions, setting strict constraints (e.g., "use only retrieved documents"), and incorporating Chain-of-Thought reasoning to guide the model toward logical, fact-based response generation. 3. Establish robust Post-Processing Verification mechanisms, including the application of automated contextual grounding guardrails, confidence scoring, and human-in-the-loop oversight for critical outputs, to detect and remediate residual hallucinations prior to final presentation.