Advances in AI Spur Creation of AI-Detection Tools

Advances in AI Spur Creation of AI-Detection Tools
  • ChatGPT passed multiple high-level exams recently.
  • Americans in some industries are afraid they could lose their jobs to AI.
  • There’s now a demand for tools that will detect AI.
  • OpenAI recently released a free web tool to do just that.

(NewsReady.com) – In the 1960s, the cartoon “The Jetsons,” showed kids what the world could look like in the future. The television show depicted a lot of artificial intelligence (AI), even though it wasn’t called that back then. For example, the characters were able to press a couple of buttons and receive their dinner.

Sixty years after the cartoon debuted, it’s becoming more clear AI is the future. There have been significant advances in the field. That is spurring the creation of tools to detect when the technology is used.

ChatGPT Advances

A program known as ChatGPT has been in the headlines a lot lately. The AI program made history when it passed all three parts of the US medical licensing exam. Professors at the University of Minnesota Law School also tested the program to see if it could pass four exams and it did. The bot reportedly received a C+ average on all of the tests.

Christian Terwiesch, a professor at the University of Pennsylvania’s Wharton School of Business, the school former President Donald Trump attended, recently wrote a paper detailing ChatGPT’s performance during a business management exam. The program performed very well and had no problem passing it.

The news has caused concern that people might use AI programs. Teachers are particularly worried about students turning in assignments that they generated by such programs. That has led to the development of AI-detection tools.

AI Detection Tools

OpenAI, the creator of ChatGPT, created a web-based tool to help educators and others detect when people are using the programs to write content. Jan Leike, OpenAI head of alignment, told Axios that the new tool has “both false positives and false negatives.”

According to Axios, users will be able to use the tool by placing blocks of text into the system in order to rate based on how likely it was that it was generated by an AI system. The tool has a five-point scale, rating the content likely, unlikely, unclear, possible, or AI-generated.

A startup known as Fictitious.ai is also testing out a program that could be used by schools to figure out whether a student has used such a program to do their work. The tool would analyze each paragraph rather than the entire body of work, then provide a score afterward.

Copyright 2023, NewsReady.com