OpenAI launches ChatGPT “nemesis”: it can identify whether articles are ghostwritten by artificial intelligence

Morning news on February 1, Beijing time, it was reported that the American artificial intelligence company OpenAI had launched the popular chat or writing robot product ChatGPT in the past. Today, the company announced in an official blog post that it has launched a new tool that can identify whether some text is written by artificial intelligence software.

The tool ChatGPT was launched in November last year and immediately became popular on the Internet and was praised and welcomed by users. Users can enter some prompt text, and this tool can write various articles, essays, jokes, and even poetry. However, this artificial intelligence writing software has also raised questions about copyright and plagiarism.

OpenAI's newly launched product is called "AI classifier". The language model used by this software has been trained on a large number of human manuscripts and artificial intelligence manuscripts on the same topic, and can identify whether a certain manuscript is Written by humans or generated by artificial intelligence software.

OpenAI said that this product integrates technology from multiple suppliers and can be used to solve various real-world problems, such as software robots spreading false information crazily on social media and misconduct in writing academic papers.

This product is still in the beta stage. OpenAI stated that the recognition performance of the tool is not very stable for articles less than 1,000 characters. In addition, articles written by artificial intelligence can deceive AI classifiers after being edited by humans.

The company said it will let the public use this new product, hoping to get some feedback from users on whether such an imperfect artificial intelligence product can produce some practical uses.

The company said that in the school education industry, the development team realized that identifying articles ghostwritten by students using artificial intelligence software is a very important topic. It is also important to understand what the AI ​​classifier has to do with identifying student articles. deficiencies and what impact they will have.

Since ChatGPT became popular on the Internet and received likes from millions of users, it has triggered alarm in some school districts in the United States. School districts, including New York City, have announced that they will ban students from using this artificial intelligence writing tool. They are worried that students will use this tool to write articles to deceive teachers or commit plagiarism.

In order to help teachers and schools identify whether students' articles are written by artificial intelligence, identification tools developed by third parties have emerged in the industry, including "GPTZeroX".

OpenAI stated that they are actively communicating with the education industry to discuss the uses or limitations of ChatGPT, and will continue to develop and improve tools for identifying articles written by artificial intelligence.