Summary of Using Machine Learning to Distinguish Human-written From Machine-generated Creative Fiction, by Andrea Cristina Mcglinchey and Peter J Barclay
Using Machine Learning to Distinguish Human-written from Machine-generated Creative Fiction
by Andrea Cristina McGlinchey, Peter J Barclay
First submitted to arxiv on: 15 Dec 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Machine Learning (cs.LG)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This research paper tackles the issue of detecting deceptive text created by Large Language Models (LLMs) in creative writing. The rise of LLMs like ChatGPT has raised concerns about the potential impact on literary culture, as AI-generated content may compromise the quality and originality of published works. To address this problem, the authors trained machine learning classifier models to distinguish human-written from machine-generated short stories, focusing on classic detective novels. The results show that a Naive Bayes and Multi-Layer Perceptron classifier achieved an accuracy of over 95%, outperforming human judges with an accuracy of less than 55%. This approach worked well with short text samples (around 100 words), which is challenging to classify. The authors have also developed an online proof-of-concept classifier tool, AI Detective, as a first step towards creating lightweight and reliable applications for editors and publishers to protect the economic and cultural contribution of human authors. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This study is about finding ways to tell if someone has used a computer program to write creative stories. With new AI tools that can generate texts, some people are worried that this could harm literary culture by making it hard for real writers to get published. The researchers trained special machines to look at short stories and figure out which ones were written by humans and which ones were made by computers. They found that the machines were really good at telling the difference (over 95% accurate!), much better than people themselves. This could help editors and publishers make sure they’re not publishing fake work, and protect the jobs of real writers. |
Keywords
» Artificial intelligence » Machine learning » Naive bayes
