OpenAI introduces invisible watermark for texts

OpenAI has announced that it will now add an invisible watermark to texts generated with the AI models ChatGPT and Codex. This measure is specifically aimed at users within the European Union and is intended to comply with new regulations regarding the use of artificial intelligence.

The watermark is automatically embedded in the generated content and is designed to verify the origin of the texts without being visually noticeable to the user. This approach is important in the context of increasing concerns about misinformation and copyright, where it is essential to trace where certain content comes from and who holds the responsibilities.

The implementation of the watermark is a step by OpenAI to promote transparency and accountability in the use of AI-generated content. It is a response to the necessity for responsible use of AI tools within the spectrum of digital ethics, and a way to ensure that such technologies are applied safely and responsibly.

With this new feature, OpenAI hopes not only to meet legal requirements but also to strengthen users' trust in the technology by providing them with a way to ensure the authenticity of the generated texts. It is a signal to the rest of the industry about the need for transparency in AI developments.

Read the full article from Bright.