AI Search Just Crossed the Credibility Threshold: OpenAI, Perplexity, and the 5% Hallucination Benchmark

AI Search Just Crossed the Credibility Threshold: OpenAI, Perplexity, and the 5% Hallucination Benchmark

导读：OpenAI、Perplexity 与斯坦福 HAI 实验室的新一轮发布，将 AI 搜索的可信度之争推向了临界点。表面上看，综合幻觉率降至 4.8% 似乎宣告着“信任高原”的到来，但一线从业者尖锐指出：不具对抗性的测试集、静默的时效性问题以及检索管道中的结构性文盲，正在让