Tuesday, the creators of a ChatGPT bot that caused a sensation due to its ability to imitate human writing unveiled a program aimed to determine whether written works were created by artificial intelligence.
Concerns that the software could be used to aid students with homework and help them cheat on examinations sparked heated debate at schools and colleges in the United States and throughout the world prior to the disclosure.
OpenAI, a company based in the United States, announced in a blog post on Tuesday that its detection tool has been trained to discern between human- and AI-authored content.
The OpenAI bot, which recently received a major infusion of funding from Microsoft, reacts to simple commands with reams of text derived from Internet data.
OpenAI warned that its tool is susceptible to error, especially with texts having fewer than one thousand characters.
“Although it is impossible to reliably detect all AI-written text, we believe that good classifiers can mitigate false claims that AI-generated text was written by a human,” OpenAI wrote in a blog post.
“For instance, conducting automated disinformation campaigns, employing AI tools for academic dishonesty, and posing an AI chatbot as a human.”
Last Monday, a prestigious French university prohibited students from using ChatGPT to do homework, the first prohibition of its kind in the country.
The decision was made shortly after it was revealed that ChatGPT had passed tests at a U.S. law school after composing essays on subjects ranging from constitutional law to taxation.
ChatGPT continues to make factual errors, but educational institutions have moved swiftly to ban the AI tool.
“We recognize that identifying AI-written text has been a topic of discussion among educators, and it is equally important to recognize the limitations and effects of AI-generated text classifiers in the classroom,” OpenAI noted in a blog post.
“We are engaging with educators in the United States to learn about their classroom experiences and to discuss ChatGPT’s capabilities and limitations.”
New York and other jurisdictions have prohibited its usage in schools.
A number of Australian colleges have stated that they will alter their exam formats to ban artificial intelligence (AI) tools and consider their use to be cheating.
OpenAI suggests using the classifier solely with English text, as its performance in other languages is inferior.
»ChatGPT developer develops AI-detection tool«