ChatGPT's secret AI content detector

OpenAI has announced its plan to add an invisible watermark to texts generated by ChatGPT and Codex, with the goal of making AI-created content easier to identify. This initiative will begin in Europe over the coming weeks, in accordance with the European AI Act. The new technology, called textGrain, involves embedding an unseen statistical marker into specific words during the text generation process. This allows for detection when analyzing the text using specialized tools. The watermark aims to improve the ability to distinguish artificial content from human-generated, although its effectiveness may be affected by short texts, editing, or paraphrasing. The company also emphasizes that the system cannot determine human involvement, identify the source of the content, or verify whether it contains accurate information. The initial application of this technology will be in Europe, in line with the new regulations.

articleDetails.fontSize.sizeحجم النص

تنويه: هذا ملخص تم إنشاؤه بواسطة الذكاء الاصطناعي

عرض المقال بالكامل