OpenAI announced the release of a tool for recognizing AI-crafted text, such as ChatGPT, created by its own AI.
ChatGPT has earned much attention in academia, as students rely on it to fulfil various duties, similar to a virtual assistant. Unfortunately, some are taking advantage of this technology and using it to fabricate content that they are attempting to submit as their own–essentially, cheating. The difficulty lies in figuring out which answers were written by a human and which by an AI.
Here Is How To Detect ChatGPT Content
OpenAI’s Classifier tool
Login
Copy your content
Paste it in OpenAI’s Classifier tool and Submit
OpenAI noted that there are a few points to bear in mind with the classifier. Primarily, the greater the number of words, the more reliable the tool. Anything below 1,000 characters and the accuracy may be reduced.
Additionally, the classifier only applies to the English language and cannot detect AI-written code. Lastly, there is a possibility that AI-generated text may be edited in a way to bypass the classifier.