Ghostbuster is a model-agnostic tool to detect AI-generated text. It offers a unique approach to distinguishing between human-written and AI-generated content. Here’s a detailed description of what Ghostbuster provides:
Key Features of Ghostbuster:
- AI-Generated Text Detection: Ghostbuster specializes in identifying whether a document has been generated by AI. It works by analyzing the text and determining its likelihood of being AI-created.
- Structured Search and Analysis: The tool passes documents through a series of weaker language models and conducts a structured search over possible combinations of their features.
- Training and Classification: Ghostbuster is trained on a dataset that includes news, student essays, and creative writing, predominantly in British and American English. It uses this training to classify the target document as AI-generated or not.
- Reliability for Different Text Lengths: The tool is known to be less reliable for texts under 100 words and has known limitations for texts between 100 and 250 words.
- Ghostbuster’s predictions are less reliable for shorter texts, texts in domains far from its training data, non-English languages, and texts written by non-native English speakers.
- It may also struggle with AI-generated text that has been edited or paraphrased by humans.
Ideal for:
- Educators and academic institutions looking to identify AI-generated student essays.
- Content creators and publishers seeking to verify the authenticity of written content.
- Organizations and individuals interested in distinguishing between human-written and AI-generated texts.
