Advanced AI detection tools don’t treat a document as a single giant block of text. Rather, these tools break down the text into segments, sentences, and paragraphs. These individual chunks are scored as segments. If you get a 30% score on a 10-page paper, it likely means that roughly three pages worth of text display the patterns that are typical of an LLM. If you are using an AI detector and a particular document gets a low score (e.g., 30% AI-generated), this usually indicates a hybrid document. A hybrid document is usually written by a human and then edited with the assistance of an AI tool or the AI-generated text is added on. What does your AI detection score mean?

A false positive happens when a detector mistakenly flags human work as AI-generated. Overall, we measure our false positive rate to be approximately 1 in 10,000: sometimes a bit higher, or a bit lower, depending on the kind of writing and other variables. While Pangram is still relatively reliable across all domains, you can be more confident in Pangram's accuracy when the text is long, in complete sentences, and requires more original input from the writer. All About False Positives in AI Detectors

For this reason, we recommend against screening things like short bullet point lists and outlines, math, very short (e.g. single sentences) responses, and extremely formulaic text such as long lists of data, spreadsheets, template based writing, and instruction manuals. All About False Positives in AI Detectors

Pangram Text is designed to detect AI-generated content with a near-zero false positive rate. Our rigorous training approach minimizes errors and allows the model to detect AI text by analyzing and understanding subtle cues in the writing. We train an initial model on a small but diverse dataset of approximately 1 million documents comprised of public and licensed human-written text. The dataset also includes AI-generated text produced by GPT-4 and other frontier language models. How AI Detection Works