推文对比了负责任的AI用法(如使用高级模型、人工复核引用)与行业现状。文章引用《纽约杂志》报道,指出媒体创业者Steven Rosenbaum在其著作《The Future of Truth》中被发现包含多条错误归因或伪造的引文,疑似由AI生成,凸显了非虚构出版业对模型幻觉的脆弱性。问题核心在于,行业长期缺乏确保事实准确的合同义务与系统性核查流程。业内人士表示漏洞普遍存在,许多AI生成的内容正被当作原创作品使用。
This is a skill &; having standards issue.
I use AI in book writing. But I write the full draft of the text myself (using AI for feedback or when stuck to give options), use advanced models that hallucinate rarely, read every reference AI finds, &; have multiple models check work.