创意写作为什么被 Turnitin 误判为 AI

创意写作可能因 Turnitin 文档记录的几个原因触发误报。模型分析的是它所说的"qualifying text"（符合条件的文本），只包含用标准语法句子写的散文。一篇 Turnitin 帮助文章说模型"does not reliably detect AI-generated text in the form of non-prose, such as poetry, scripts, or code." 短的创意作品还有额外问题：FAQ 说在几百词的文档中"the prediction will be mostly 'all or nothing' because we're predicting on a single segment without the opportunity to overlap." FAQ 还列出了产生误报的特征，包括"content without a lot of structural variation"和"text that literally repeats itself." 使用重复母题或统一句式的创意写作即使完全由人写，也可能匹配这个特征。

FAQ 定义了模型检查的范围：

"This qualifying text includes only prose sentences, meaning that we only analyze blocks of text that are written in standard grammatical sentences and do not include other types of writing such as lists, bullet points (short non-sentence structures), or other non-sentence structures."

下一句："This percentage is not necessarily the percentage of the entire submission."

一篇 Turnitin 帮助文章提供了模型不能可靠检测的更完整列表：

"The model does not reliably detect AI-generated text in the form of non-prose, such as poetry, scripts, or code, nor does it detect short-form/unconventional writing such as bullet points, tables, or annotated bibliographies."

这对创意写作作业很重要。如果你的作品包含诗歌片段、剧本式对话或非常规格式，那些部分可能根本不被分析。你收到的分数只反映符合条件的散文部分，这会在你写的内容和分数代表的含义之间制造差距。

创意写作作业通常很短。一篇 500 词的微小说或 300 词的速写落入了模型挣扎的范围。FAQ 说：

"In shorter documents where there are only a few hundred words, the prediction will be mostly 'all or nothing' because we're predicting on a single segment without the opportunity to overlap."

下一句："This means that some text that is a mix of AI-generated and original content could be flagged as entirely AI-generated."

2023 年 12 月的一条 release note 补充："Results show that our accuracy increases with a little more text so submissions that contain less than 300 words may result in an AI writing score that is likely less accurate."

如果你的创意作品不到 300 词，分数可能不太准确。如果不到几百词，预测可能摆向极端，把混合或完全人写的内容标记为全部 AI 生成。Turnitin 报告显示 100% AI，是怎么来的按片段追了一遍这个摆动。

FAQ 列出了容易产生误报的文本特征：

"Sometimes false positives (incorrectly flagging human-written text as AI-generated), can include content without a lot of structural variation, text that literally repeats itself, or text that has been paraphrased without developing new ideas."

下一句："If our indicator shows a higher amount of AI writing in such text, we advise you to take that into consideration when looking at the percentage indicated."

这些特征中有几个与常见创意写作技巧重合。使用刻意重复作为风格手段的作品可能匹配"text that literally repeats itself." 具有一致叙事声音和统一句子节奏的作品可能匹配"content without a lot of structural variation." 官方的建议是对百分比"take that into consideration"而非把它当作确定性的。单个的风格习惯也会被这样读，破折号是 AI 写作的马脚吗问的就是这件事。

检测模型通过把文本切成重叠片段并给每个分配概率分来处理：

"When a paper is submitted to Turnitin, sentences from the submission are extracted and segmented into overlapping sections for prediction analysis. Each segment is classified by the AI detection model and given a value between 0 and 1, denoting the probability of the text being likely human or AI-generated."

在较长的文档中，片段重叠，这意味着单个句子可以接收多个分数并汇总在一起。这种汇总会平滑掉单个预测错误。在短的创意作品中，可能只有一个片段没有重叠，这就是预测变成全有或全无的原因。每个片段是按什么打分的，见困惑度与突发性之外，AI 检测器还在测什么。

总结一下我们讲的内容：

• Turnitin 的模型只分析符合条件的散文句子。诗歌、剧本和非散文格式可能不被分析。
• 一篇 Turnitin 帮助文章列出了诗歌、剧本和注释书目为模型不能可靠检测的文本类型。
• 几百词的短文档得到全有或全无的预测，没有重叠片段来平滑错误。
• 2023 年 12 月的 release note 指出不到 300 词的提交可能不太准确。
• FAQ 建议教师对结构变化少或自我重复的文本"take that into consideration"。
• 刻意重复和一致声音等创意写作技巧可能匹配误报易发特征。

如果收到创意写作作品的 Turnitin 报告并想处理被标记的段落，可以导入报告处理。符合条件时可以免费继续降 AI。