Making ChatGPT do all your homework? OpenAI brings new tool to detect AI-generated text

ChatGPT's creator OpenAI has taken a big step towards countering the issue of copying AI-generated text. The new tool has been named AI Text Classifier. The creators trained the classifier to distinguish between text written by a human and text written by AIs. While the new tool has a lot of promise, the creators at OpenAI acknowledge that it may not be perfect.

The creators claim that it is impossible to reliably detect all AI-written text, but the AI Classifiers can provide helpful information against false claims that AI-generated text was written by a human for example, running automated misinformation campaigns, using AI tools for academic dishonesty, and positioning an AI chatbot as a human.

They have clearly stated that the classifier is not fully reliable. According to the company blog, the classifier correctly identifies 26 percent of AI-written text (true positives) as “likely AI-written,” while incorrectly labeling human-written text as AI-written 9 percent of the time (false positives). The company's classifier’s reliability typically improves as the length of the input text increases. However, the creators claim that the new classifier is significantly more reliable on text from more recent AI systems.

Also read: 'Chinese ChatGPT' to launch soon to take on OpenAI and Google

The classifier is publicly available and users can offer feedback. The "imperfect tool" for detection of AI-generated text will continue to get better according to the creators at OpenAI.

A tool for Educators

Educators all over the world have shown growing concerns about the new ChatGPT AI and its capability to offer high-quality, human-like text. The AI was not only able to pass exams for lawyers and managers but also medical professionals. The creators have also acknowledged this issue.

The creators state, "We recognize that identifying AI-written text has been an important point of discussion among educators, and equally important is recognizing the limits and impacts of AI generated text classifiers in the classroom. We have developed a preliminary resource on the use of ChatGPT for educators, which outlines some of the uses and associated limitations and considerations. While this resource is focused on educators, we expect our classifier and associated classifier tools to have an impact on journalists, mis/dis-information researchers, and other groups."

Also read: Google's new AI can create various genres of music from text description

Limitations of new AI Text Classifier

• The classifier is very unreliable on short texts (below 1,000 characters). Even longer texts are sometimes incorrectly labeled by the classifier.
• Sometimes human-written text will be incorrectly but confidently labeled as AI-written by our classifier.
• We recommend using the classifier only for English text. It performs significantly worse in other languages and it is unreliable on code.
• Text that is very predictable cannot be reliably identified. For example, it is impossible to predict whether a list of the first 1,000 prime numbers was written by AI or humans, because the correct answer is always the same.
• AI-written text can be edited to evade the classifier. Classifiers like ours can be updated and retrained based on successful attacks, but it is unclear whether detection has an advantage in the long-term.
• Classifiers based on neural networks are known to be poorly calibrated outside of their training data. For inputs that are very different from text in our training set, the classifier is sometimes extremely confident in a wrong prediction.