Crux: OpenAI has released a free web-based tool to distinguish between text written by a human and AI after intense debate on whether the artificial intelligence chatbot is a tool to enhance learning or a threat or both.
AI detection tool: ChatGPT creator OpenAI has released a free tool designed to help detect if a particular text was written by a human or a machine. It can detect texts generated by AI from a variety of sources including ChatGPT.
The new tool, known as a classifier, was launched after weeks of discussion at schools and colleges over fears that the use of ChatGPT could fuel academic dishonesty and pose a threat to learning.
ChatGPT is an AI-powered chatbot allowing users to hold human-like conversations with an AI, answering follow-up questions, challenging incorrect premises, and rejecting inappropriate requests. GPT stands for Generative Pre-trained Transformer, a language processing model developed by the American artificial intelligence company OpenAI.
Tool not fully reliable: OpenAI has cautioned that the beta version of the tool for detecting AI-written text is imperfect and can be wrong at times. It warns the classifier is “not fully reliable” and “should not be used as a primary decision-making tool” to determine the authorship of a document.
OpenAI also says that in its tests the tool identified AI-written text as “likely AI-written” 26% of the time and gave false AI detections 9% of the time, outperforming its previous tool for AI-detected text.
How it works?
- The tool is relatively simple. You just paste the text into a box, click submit, and the system will rate how likely the text it is that has been generated by an AI system.
- It rates the text on a five-point scale: Very unlikely to have been AI-generated, unlikely, unclear, possible or likely.
Current limitations:
The company has listed several limitations such as:
- The AI classifier requires a minimum of 1,000 characters (approximately 150 – 250 words).
- The classifier isn’t always accurate. Sometimes, it can mislabel both AI-generated and human-written text.
- AI-generated text can be edited easily to evade the classifier.
- It works best in English, with performance significantly worse in other languages.
Since its launch in November, the artificial intelligence program has sparked concerns in sectors such as education and journalism over its ability to produce essays, research papers, and other written material.
Some even see ChatGPT as a direct threat to Google’s search engine. Gmail creator Paul Buchheit tweeted that the search engine “may be only a year or two away from total disruption”.
One thing that few people remember is the pre-Internet business that Google killed: The Yellow Pages!
— Paul Buchheit (@paultoo) December 1, 2022
The Yellow Pages used to be a great business, but then Google got so good that everyone stopped using the yellow pages.
AI will do the same thing to web search
Microsoft recently confirmed an extension of its partnership with OpenAI, investing $10 billion investment.
