Teachers rejoice! ChatGPT creators have released a tool to help detect AI-generated writing

Teachers rejoice! ChatGPT creators have released a tool to help detect AI-generated writing
OpenAI CEO Sam Altman recently said the worst-case scenario for AI is 'lights out for all of us.'Lucy Nicholson/Reuters
  • OpenAI has finally released a way to detect AI-generated writing from its chatbot, ChatGPT.
  • The new program comes amid concerns that AI chatbots could make it easier for students to cheat.

OpenAI has heard everyone's concerns, and it's finally working on detecting AI-generated writing.

Since its launch, the company's explosive AI chatbot, ChatGPT, has caused shockwaves in many industries due to its skilled writing and coding abilities. The bot has already accomplished several impressive feats, including passing the US Medical Licensing Exam, a Wharton MBA exam, and 4 law school courses.

However, ChatGPT has also caused handwringing among teachers and other education professionals who say the bot will help students get better at cheating and plagiarism.

Complimentary Tech Event
Transform talent with learning that works
Capability development is critical for businesses who want to push the envelope of innovation.Discover how business leaders are strategizing around building talent capabilities and empowering employee transformation.Know More

On Tuesday, the company launched a web-based program called "AI Text Classifier" to tackle that issue.

The program will flag pasted-in text with the following labels: "very unlikely," "unlikely," "unclear if it is," "possibly," or "likely" AI-generated.


OpenAI admits its tool isn't quite perfect yet: It requires a minimum of 1,000 characters to determine whether text is AI-generated and is prone to making errors.

"These tools will produce both false negatives, where they don't identify AI-generated content as such, and false positives, where they flag human-written content as AI-generated. Additionally, students may quickly learn how to evade detection by modifying some words or clauses in generated content," the company said in a blog post.

OpenAI did not immediately respond to Insider's request for comment.

OpenAI isn't the first to attempt an antidote to the AI-writing conundrum. Plagiarism detector Turnitin told Insider they are working on a similar product to detect AI-generated text from ChatGPT, which a Turnitin executive called a "mad-lib machine."

A 22-year-old college student at Princeton University also released a program to detect AI writing earlier this month called GPTZero.