ChatGPT Detector: These tools detect AI-written texts

Using a so-called AI detector, you can recognise content that has been created by ChatGPT or other artificial intelligences. There are now several free programmes for this purpose.

KI-Detector: These tools help you to detect content created by ChatGPT.

OpenAI’s language model supports many people in different fields. ChatGPT is a powerful tool that can do a wide variety of jobs. For this very reason, it is becoming increasingly important to detect content created by such chatbots. In the following, we present five free detectors:

• Botometer: This is a tool from the Information Sciences Institute at the University of Southern California. Using machine learning algorithms, it detects whether a Twitter account, a bot or a human created the content. It uses factors such as the frequency of tweets, the language used and the use of certain hashtags or URLs. It also calculates the probability of whether the content came from ChatGPT or another language model.
• Bot Sentinel: This tool can also detect Twitter accounts that exhibit behaviour similar to a bot. It examines similar parameters to Botometer and identifies accounts based on the frequency of tweets, their similarity to other Twitter bots and the use of certain keywords or hashtags.
• Botcheck.me: This is another tool that analyses Twitter accounts. It analyses parameters such as the frequency and timing of tweets as well as the similarity with other Twitter bots and the occurrence of certain hashtags or URLs.
• GPT-3 Detector: This is OpenAI’s own tool to check the content of language models. It can detect whether a text comes from their own AI or was created by a human. However, it is not specifically designed to detect ChatGPT. It can still be useful for identifying texts generated by a language model. The tool uses a deep neural network to analyse the language in the text and compare it to known patterns of human speech and GPT-3.
• Hugging Face Transformers: The company offers both a free and open source library of natural language processing tools. Among them are several models that can recognise speech-based chatbots. The pipeline tool is able to recognise whether a text was generated by ChatGPT or another language model. Hugging Face Transformers is used by researchers, developers and other professionals working with natural language processing and machine learning.