Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
41 tokens/sec
GPT-4o
59 tokens/sec
Gemini 2.5 Pro Pro
41 tokens/sec
o3 Pro
7 tokens/sec
GPT-4.1 Pro
50 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

On the Possibilities of AI-Generated Text Detection (2304.04736v3)

Published 10 Apr 2023 in cs.CL, cs.AI, and cs.LG

Abstract: Our work addresses the critical issue of distinguishing text generated by LLMs from human-produced text, a task essential for numerous applications. Despite ongoing debate about the feasibility of such differentiation, we present evidence supporting its consistent achievability, except when human and machine text distributions are indistinguishable across their entire support. Drawing from information theory, we argue that as machine-generated text approximates human-like quality, the sample size needed for detection increases. We establish precise sample complexity bounds for detecting AI-generated text, laying groundwork for future research aimed at developing advanced, multi-sample detectors. Our empirical evaluations across multiple datasets (Xsum, Squad, IMDb, and Kaggle FakeNews) confirm the viability of enhanced detection methods. We test various state-of-the-art text generators, including GPT-2, GPT-3.5-Turbo, Llama, Llama-2-13B-Chat-HF, and Llama-2-70B-Chat-HF, against detectors, including oBERTa-Large/Base-Detector, GPTZero. Our findings align with OpenAI's empirical data related to sequence length, marking the first theoretical substantiation for these observations.

PDF HTML Abstract
Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize File Document Download Save Streamline Icon: https://streamlinehq.com PDF File Document Download Save Streamline Icon: https://streamlinehq.com Markdown Bookmark Chat Bubble Oval Streamline Icon: https://streamlinehq.com Chat (Pro)
Definition Search Book Streamline Icon: https://streamlinehq.com
References (49)
  1. AI-based Text Detection (zerogpt). AI Text Detector, Jan a. URL https://www.zerogpt.com.
  2. Ai-based text detection (openai). AI Text Detector, Jan b. URL https://openai.com/blog/new-ai-classifier-for-indicating-ai-written-text.
  3. Scott Aaronson. My Projects at OpenAI, Nov 2022. URL https://scottaaronson.blog/?p=6823.
  4. Do as i can, not as i say: Grounding language in robotic affordances, 2022.
  5. Natural language watermarking: Design, analysis, and a proof-of-concept implementation. In Proceedings of the 4th International Workshop on Information Hiding, IHW ’01, pp.  185–199, Berlin, Heidelberg, 2001. Springer-Verlag. ISBN 3540427333.
  6. On the opportunities and risks of foundation models, 2022.
  7. Language models are few-shot learners, 2020.
  8. Rank-lime: Local model-agnostic feature attribution for learning to rank, 2022.
  9. Thomas M Cover. Elements of information theory. John Wiley & Sons, 1999.
  10. Prithiviraj Damodaran. Parrot: Paraphrase generation for nlu., 2021.
  11. Amit Dhurandhar. Auto-correlation dependent bounds for relational data. In Proc. of the 11th Workshop on Mining and Learning with Graphs. Chicago, 2013.
  12. Tom Fawcett. An introduction to roc analysis. Pattern recognition letters, 27(8):861–874, 2006.
  13. Gltr: Statistical detection and visualization of generated text, 2019.
  14. Latent dirichlet allocation (lda) and topic modeling: models, applications, a survey, 2018.
  15. Self-attentive sequential recommendation, 2018.
  16. Dense passage retrieval for open-domain question answering, 2020.
  17. Ask me what you need: Product retrieval using knowledge from gpt-3, 2022.
  18. A watermark for large language models, 2023.
  19. Paraphrasing evades detectors of ai-generated text, but retrieval is an effective defense, 2023.
  20. Deep learning–based social media misinformation detection. IEEE Software, 39(1):53–59, 2022. doi: 10.1109/MS.2022.3053106.
  21. Watermarking digital image and video data. a state-of-the-art overview. IEEE Signal processing magazine, 17(5):20–46, 2000.
  22. Detecting fake content with relative entropy scoring. In Pan, 2008.
  23. Lucien Le Cam. Asymptotic methods in statistical decision theory. Springer Science & Business Media, 2012.
  24. Gpt detectors are biased against non-native english writers, 2023.
  25. William Lifferth. Fake news, 2018. URL https://kaggle.com/competitions/fake-news.
  26. Topics as entity clusters: Entity-based topics from language models and graph neural networks, 2023.
  27. Learning word vectors for sentiment analysis. In Proceedings of the 49th annual meeting of the association for computational linguistics: Human language technologies, pp.  142–150, 2011.
  28. Natural language watermarking via morphosyntactic alterations. Comput. Speech Lang., 23(1):107–125, jan 2009. ISSN 0885-2308. doi: 10.1016/j.csl.2008.04.001. URL https://doi.org/10.1016/j.csl.2008.04.001.
  29. Detectgpt: Zero-shot machine-generated text detection using probability curvature, 2023.
  30. Don’t give me the details, just the summary! topic-aware convolutional neural networks for extreme summarization. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, pp.  1797–1807, Brussels, Belgium, October-November 2018. Association for Computational Linguistics. doi: 10.18653/v1/D18-1206. URL https://aclanthology.org/D18-1206.
  31. OpenAI. Ai text classifier. https://platform.openai.com/ai-text-classifier, 2023.
  32. OpenAI. Gpt-4 technical report, 2023.
  33. Question answering survey: Directions, challenges, datasets, evaluation matrices, 2021.
  34. Information theory: From coding to learning, 2022.
  35. SQuAD: 100,000+ questions for machine comprehension of text. In Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, pp.  2383–2392, Austin, Texas, November 2016. Association for Computational Linguistics. doi: 10.18653/v1/D16-1264. URL https://aclanthology.org/D16-1264.
  36. A neural attention model for abstractive sentence summarization. In Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pp.  379–389, 2015.
  37. Can ai-generated text be reliably detected? arXiv preprint arXiv:2303.11156, 2023.
  38. Tyler Schildhauer. Fake news detection in the era of ai. In Proceedings of the 25th ACM Conference on Computer-Supported Cooperative Work and Social Computing, pp.  1–10. ACM, 2022. doi: 10.1145/1234567.1234567.
  39. Building end-to-end dialogue systems using generative hierarchical neural network models. In Proceedings of the Thirtieth AAAI Conference on Artificial Intelligence, 2015.
  40. Multilingual language generation and automatic writing evaluation with transformer models. arXiv preprint arXiv:2104.06399, 2021.
  41. Release strategies and the social impacts of language models, 2019.
  42. Salil Pravin Vadhan. A study of statistical zero-knowledge proofs. PhD thesis, Massachusetts Institute of Technology, 1999.
  43. Attention is all you need. Advances in neural information processing systems, 30:5998–6008, 2017.
  44. Robustness of the digital image watermarking techniques against brightness and rotation attack, 2009.
  45. A comprehensive review on digital image watermarking, 2022.
  46. Zhen Wang. Modern question answering datasets and benchmarks: A survey, 2022.
  47. Larry Wasserman. Lecture notes for stat 705: Advanced data analysis. https://www.stat.cmu.edu/~larry/=stat705/Lecture27.pdf, 2013. Accessed on April 9, 2023.
  48. Cross-lingual phrase retrieval, 2022.
  49. Xiaofei Zou and Xu Ling. Ai-based detection of misinformation in social media. IEEE Access, 9:112408–112418, 2021. doi: 10.1109/ACCESS.2021.3104419.
User Edit Pencil Streamline Icon: https://streamlinehq.com
Authors (6)
  1. Souradip Chakraborty (36 papers)
  2. Amrit Singh Bedi (75 papers)
  3. Sicheng Zhu (15 papers)
  4. Bang An (33 papers)
  5. Dinesh Manocha (366 papers)
  6. Furong Huang (150 papers)
Citations (106)
View on Semantic Scholar