精选理由
RadLE 2.0测试发现AI读X光常常自信地犯错，人类医生还远领先。搞医疗AI的该看看这个教训。
RadLE 2.0基准测试评估AI模型在放射学中判断何时应由人类接手诊断的能力。测试发现，许多AI模型在给出错误结论时仍表现出完全自信，而人类放射科医生的准确率仍显著更高。该基准聚焦于模型的自知力，强调AI必须学会在不确定时保持沉默，才能安全投入临床。
原文 · Decoder
AI chatbots reading X-rays can be dangerously confident even when they're wrong
The RadLE 2.0 benchmark tests whether AI models in radiology can tell when they should leave a diagnosis to a human. Many models deliver wrong findings with full confidence, and human radiologists are still well ahead. Before AI can diagnose on its own, it needs to learn when it's better to say nothing. The article AI chatbots reading X-rays can be dangerously confident even when they're wrong appeared first on The Decoder .