With the popularity of ChatGPT, it is becoming increasingly difficult to distinguish human-generated content from machine-generated content. To reduce this problem, OpenAI has recently released an open source tool for AI-generated text detection.
LIA is becoming an increasingly popular content generation tool, especially with the popular ChatGPT. This can be articles, social media posts, and even college theses. Many institutions have rightly complained about this trend, with some students using AI as a shortcut.
Open AI: An AI discovery tool with limitations
Following the controversy surrounding the use of ChatGPT to generate academic texts, OpenAI decided to launch a new tool called “AI Text Classifier”.. This allows users Find out if the text was written by a human or by an artificial intelligence. The tool uses a machine learning system that analyzes text and classifies it into different categories. Results vary between It may have been created by artificial intelligence. And “Very unlikely”.
However, OpenAI warns against this Errors are possible And he recommends not to rely solely on this score. Teachers can use the tool for the information base, however The results should not be considered a reliable reference. It is always important to include human judgment in this process.
“We don’t really recommend using this tool as a sole reference. We know it can be wrong,” said Lama Ahmed, Director of Search Strategy at Lama Ahmed. Open AI.
Before the tool was launched, a demo was done with CNN on ChatGPT-generated scripts. In this demo, it was unlikely that the AI would have produced a clip from the book “Peter Pan”. However, in a blog post published by OpenAI, it is mentioned that in 5% of cases an error is made when classifying human texts as AI-produced.
ChatGPT: Disturbing use in education
ChatGPT, which was released last November, has become the most widely used tool on the Internet. This is due to its ability to accurately answer specific questions. At the same time, it allows to write posts, stories and other materials with great ease. However, the chatbot Some educators are concerned For its ability to mimic human handwriting. Most worry The use of artificial intelligence can harm learning. In fact, students may be tempted to copy the work generated by the model without understanding the subject in depth. Additionally, use ChatGPT to generate responses May question the correctness of the teacher’s assessment. They will find it difficult to determine whether the student has already acquired the required knowledge.
As a result, companies like OpenAI and Turnitin are working on it ChatGPT created content detection tools. These tools will allow educators to identify model-generated works. Ultimately, the concern about ChatGPT points to the need to better understand the consequences of using these technologies in education and take appropriate steps to manage them.