With the popularity of text-generation AI tools like chatGPT and GPT3/3.5, detecting AI-generated content has become important for publishers and readers.
We look at tools for detecting AI-generated content.
Table of Contents
Tools for detecting AI-generated content
HuggingFace OpenAI Detector
HuggingFace provides a free online tool for detecting AI-generated content.
It has some limitations in that it’s not good with detecting content created by GPT-3 and is not very scalable if you are a large or medium size company that does serious publishing.
OriginalityAI is a paid product for detecting AI-generated content. It is built for content publishers who have teams of writers they manage.
GPTZero is a new tool that can detect text generated by new AI tools like chatgpt (chat gpt) or GPT-3 AI-based writers. It uses variables like perplexity to measure whether a text was generated by an AI.
OpenAI AI Text Classifier
OpenAI, the creator of GPT 3.x, DALL-E, and chatGPT has released its own AI-generated text detector, called AI text Classifier. It classifies the text into three categories – Human-Written, AI_generated, and Misclassfied Human-Written.
Build your own tool for detecting AI-generated content
You could also build your own using this open-source code from OpenAI.