A new study published on arXiv explores the ability of humans to distinguish between AI-generated and human-written Japanese haiku. Researchers found that while models like LLM-JP, Gemma-2B, and LLaMA-2 showed moderate detectability, advanced models such as GPT-5 and Gemini 2.5 performed at chance levels. Interestingly, participants' aesthetic judgments of poeticness and fluency did not correlate with their ability to correctly identify the author, suggesting an attribution bias. AI
IMPACT Suggests that as LLMs improve, distinguishing AI-generated creative content from human work will become increasingly difficult.
RANK_REASON Academic paper published on arXiv detailing a study on AI-generated poetry. [lever_c_demoted from research: ic=1 ai=1.0]
- arXiv
- Gemini 2 5
- Gemma 2B
- GPT-5
- Haiku
- Hugging Face
- large-language models
- LLaMA-2
- LLM-JP
- StableLM-7B
- Tokyo
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
