ChatGPT owner launches ‘flawed’ tool to detect AI-generated text

OpenAI, the creator of the popular chatbot ChatGTP, has released a software tool to identify text generated by artificial intelligence, the company announced in a blog post on Wednesday (February 1).

ChatGPT is a free program that generates a text response to a prompt, including articles, essays, jokes, and even poetry, that has grown in popularity since its debut in November, while raising privacy concerns. authorship and plagiarism.

The AI ​​classifier, a language model trained on the dataset of human-written and AI-written text pairs on the same topic, aims to distinguish between AI-written text. It uses a variety of vendors to address issues such as automated disinformation campaigns and academic dishonesty, the company said.

In its public beta mode, OpenAI recognizes that the detection tool is very unreliable on texts under 1,000 characters, and the text written by the AI ​​can be modified to fool the classifier.

“We are making this classifier publicly available for feedback on the usefulness of flawed tools like this,” OpenAI said.

“We recognize that identifying text written by AI has been an important point of discussion among educators, and it is equally important to recognize the limitations and impacts of AI-generated text classifiers in the classroom. of class.”

Since ChatGPT debuted in November and gained popularity with millions of users, some of America’s largest school districts, including New York, have banned the AI ​​chatbot over concerns that students are using the text generator. to cheat or plagiarize.

Others have created third-party detection tools, including GPTZeroX, to help teachers detect AI-generated text.

OpenAI said it was engaging with educators to discuss the capabilities and limitations of ChatGPT, and would continue to work on detecting AI-generated text.