As the use of ChatGPT and other Large Language Models (LLMs) become mainstream tools for writing and researching, it’s important to know when something produced with the help of AI can be considered human generated or AI generated.
Originality.ai is an AI content detector and plagiarism checker. According to their website, their results against OpenAI ChatGPT-4 are 99.37% accurate with 1.56% false positives. I was curious about what Originality.ai considered “AI generated”, and they list it on their webpage:
- Al Generated and Not Edited = Al Generated Text
- Al Generated and Human Edited = Al Generated Text
- Al Outline, Human Written and heavily Al Edited = Al Generated Text
- Al Research and Human Written = Original Human-Generated
- Human Written and Edited with Grammarly = Original Human-Generated
- Human Written and Edited = Original Human-Generated
Using AI to generate text is not bad per se. The problem arises when one we present those ideas as our own. Maybe we should adopt some kind of convention to signal that a text was generated by an LLM tool.