OpenAI has now released a new tool that helps detect whether the text was written by a human or AI. OpenAI has also confirmed that the classifier isn’t always accurate. So, it can mislabel the content.
You just need to paste the text and start the test and it will tell if the text was most likely written by AI. It also requires a minimum of 1,000 characters for the input.
I tried the tool by checking on the text generated by ChatGPT and the articles of my blog. This tool confirmed that the ChatGPT text was ‘possibly AI-generated’ and classed my articles as “very unlikely AI-generated.” There could be five results: Very unlikely to have been AI-generated, unlikely, unclear, possible or likely.
OpenAI suggested not to use the tool alone to determine content’s authenticity as there may be.
After ChatGPT became available to all the people, it has raised concerns for Educational Institutes. This tool can write assignments and help with code. So, several colleges have already banned ChatGPT on their networks. Experts are also worried about the rise of AI-generated misinformation. Content writers are also using this tool to save time. So, a tool that can help in detecting AI-generated content was a need.
