OpenAI releases “not fully reliable” tool to detect AI generated content

ChatGPT has been creating waves across the internet with its writing ability and responses to requests. The use of ChatGPT has been full of     1     . There have been concerns that students can use this tool to     2     AI-generated work and claim it as their own.

OpenAI, the research laboratory behind ChatGPT, has released a tool designed to     3     whether text has been written by artificial intelligence, but warns it’s not completely reliable – yet. In a blog post on Tuesday, OpenAI linked to a new     4     tool that has been trained to distinguish between text written by a human and that written by a variety of AI, not just ChatGPT.

The tool could be useful in cases where AI was used for “    5     dishonesty” and when AI chatbots were positioned as humans, they said. But they admitted the tool “is not fully reliable” and only correctly identified 26% of AI-written English texts. It also incorrectly     6     human-written texts as probably written by AI tools 9% of the time.

“Our classifier’s reliability     7     improves as the length of the input text increases. Compared to our     8     released classifier, this new classifier is significantly more reliable on text from more recent AI systems.”

Since ChatGPT was opened up to public    9    , it has given rise to a wave of concern among educational institutions across the world that it could lead to cheating in exams or assessments. Lecturers in the UK are being urged to review the way in which their courses were     10    , while some universities have banned the technology entirely and returned to pen-and-paper exams to stop students using AI.