24% of AI-generated research documents found to be smart plagiarism, ACL 2025 paper shows
danish037 · x · 2026-09-10
A study accepted to ACL 2025, 'All That Glitters is Not Novel', documents plagiarism in AI research agents.
- 13 experts compared 50 LLM-generated research documents against prior work: 24% were paraphrased or heavily borrowed (with one-to-one methodological mapping), verified by the original papers' authors.
- The remaining 76% showed varying degrees of similarity; only a small fraction appeared genuinely novel. Generated documents neither cite sources nor trip built-in plagiarism detectors.
- Controlled experiments show automated detectors are inadequate at catching idea-level plagiarism, prompting a call for careful assessment of AI-generated research in academic publishing.
More from Research
- DexGPT: a GPT-written real2sim pipeline turns one monocular hand GIF into 44-DOF dexterous sim, open-sourced — animesh_garg · 2026-09-10
- DeepSeek V4.1 Flash is actually 748B params, safetensors analysis shows — DistanceSolar1449 · 2026-09-10
- Neural Harmonic Textures at ECCV beats all 3DGS and NeRF variants, code open-sourced — ZGojcic · 2026-09-10
- NeurIPS Learn2Design competition offers €10,000 for new optimization algorithms for physical systems — MarioKrenn6240 · 2026-09-10
- Salesforce paper: fine-tuning on evolved harnesses makes weak models worse on all 7 tasks — omarsar0 · 2026-09-10
- Gen2Balance: text-to-video synthesis fills long-tailed data for SOTA action recognition — dimadamen · 2026-09-10