ChatGPT, the artificial intelligence chatbot created by OpenAI in November 2022, is making waves in the world of education. Schools across the United States have banned ChatGPT from their networks due to concerns that students may use it to cheat on assignments and tests. The technology has also raised questions about AI-generated misinformation and bots posing as humans.
In response to these worries, OpenAI has released a free web-based tool designed to help educators and others determine if a particular chunk of text was written by a human or an AI system. The tool works best on English text samples greater than 1,000 words, though its performance is significantly worse in other languages. It also does not work for computer code written by humans or AI.
When users copy a chunk of text into the box, the system will rate how likely it is that the text was generated by an AI system using a five-point scale ranging from “very unlikely” to “likely”. OpenAI cautions that the tool is imperfect and performance varies based on how similar the text being analyzed is to the types of writing OpenAI’s tool was trained on; it can produce both false positives and false negatives.
The emergence of powerful tools like ChatGPT has caused many people to worry about potential misuse and abuse of this technology. While OpenAI's new tool may not be perfect, it could help educators better identify when students are attempting to pass off AI-generated work as their own.