PANews reported on August 5 that according to the Wall Street Journal, OpenAI is developing a novel text watermarking technology to identify and detect text generated by ChatGPT. This technology works by embedding invisible watermarks in the text, which is achieved by making slight changes to the way ChatGPT selects words, essentially creating an invisible watermark in the text that can later be detected by a separate tool.
However, there are still many challenges. For example, the recognition effect of watermarks may be greatly reduced in the face of text translation, rewriting, or insertion of special characters. At the same time, the technology may have an unfair impact on user groups that are not native English speakers. In view of these complexities, OpenAI said it will continue its research prudently and focus more on the development of audiovisual content authentication tools.
