Unverified50% confidenceFactExact time
If an AI omits critical information such as a patient's drug allergy history, past medical history, or medication dosage, LLM judges are likely to remain completely unaware and instead give high scores for accuracy and freedom from hallucinations.
1
Sources
50%
Confidence
Long-term
Relevance
9/11/2026
First Seen
Sources
Related Entities
Related Claims
Unverified对于患者端医疗AI摘要产品，缓解幻觉风险的策略包括保留原始音频供回溯验证、对数字和药名进行实体识别校验、标注低置信度段落、声明摘要仅供参考75% similarUnverifiedA study focused on AI clinical medical records found that LLM judges are good at verifying the presence of information but struggle to detect missing information, a phenomenon the researchers named Omission Blindness.74% similarUnverifiedExisting evaluation mechanisms systematically overestimate the quality of AI clinical medical records, primarily guarding against over-generation while almost never guarding against under-generation.71% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/909753API
curl https://kongchang.com/api/v1/knowledge/claims/909753MCP
get_claim(id=909753)