Paraphrasing evades detectors of AI-generated text, but retrieval is an effective defense. (arXiv:2303.13408v1 [cs.CL])

To detect the deployment of large language models for malicious use cases
(e.g., fake content creation or academic plagiarism), several approaches have
recently been proposed for identifying AI-generated text via watermarks or
statistical irregularities. How robust are these detection algorithms to
paraphrases of AI-generated text? To stress test these detectors, we first
train an 11B parameter paraphrase generation model (DIPPER) that can paraphrase
paragraphs, optionally leveraging surrounding text (e.g., user-written prompts)
as context. DIPPER also uses scalar knobs to control …

algorithms cases content creation defense deployment detect detection fake generated language language models large malicious parameter stress stress test test text train use cases