Stanford researchers have developed a tool called ‘DetectGPT’ that will help teachers identify content that has been generated using ChatGPT or other similar models. Since the popularity of ChatGPT, there have been concerns regarding students using ChatGPT to solve examination papers and write assignments. Moreover, OpenAI is also working on a tool to detect its AI-generated content.
What is DetectGPT and How Does it Work?
According to the researchers, the text generated using ChatGPT or similar models ‘occupy negative curvature regions of the model’s log probability function.’ The tool uses log probabilities to determine if the given passage has been generated using AI models; the researchers have also claimed that the model is correct 95% of the time.
Educational institutes around the world have started a crackdown on content generated using AI tools, and the students are warned of strict checking and random checks. A Bengaluru- based university has already issued an advisory to students to prevent the use ChatGPT. The institute will be conducting random checks among students, and if they fail to reproduce identical results, disciplinary actions will be taken, said the dean, Sanjay Chitnis.
OpenAI Releases Tool to Detect AI-generated Content
OpenAI has detected a tool to distinguish between human-written and AI-generated text. The spokesperson of OpenAI said; “The classifier aims to help mitigate false claims that AI-generated text was written by a human. However, it still has a number of limitations — so it should be used as a complement to other methods of determining the source of text instead of being the primary decision-making tool.”
Also read: BuzzFeed to Use ChatGPT’s Creator OpenAI to Help Create Content