<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>Detection on Dylan Ler</title><link>http://dylanler.github.io/tags/detection/</link><description>Recent content in Detection on Dylan Ler</description><generator>Hugo -- 0.133.0</generator><language>en-us</language><lastBuildDate>Sat, 18 Oct 2025 16:30:00 -0700</lastBuildDate><atom:link href="http://dylanler.github.io/tags/detection/index.xml" rel="self" type="application/rss+xml"/><item><title>Can AI Spot Its Own Kind? LLMs Detecting AI vs Human Creative Work</title><link>http://dylanler.github.io/posts/creative-authenticity-ai-vs-human-art/</link><pubDate>Sat, 18 Oct 2025 16:30:00 -0700</pubDate><guid>http://dylanler.github.io/posts/creative-authenticity-ai-vs-human-art/</guid><description>Here&amp;rsquo;s a poem. Human or AI?
The morning light falls soft on empty chairs, where conversations used to fill the air. Now silence keeps its patient, gentle watch— a house that holds the shape of those who&amp;rsquo;ve gone.
This experiment tests whether LLMs can distinguish AI-generated creative work from human work—and what their detection strategies reveal about what they consider &amp;ldquo;authentically human.&amp;rdquo;
The Experiment We curated 500 creative works:
250 human-created (published works, attributed artists) 250 AI-generated (GPT-4, Claude, Midjourney prompts) Across 5 domains:</description></item></channel></rss>