Google DeepMind introduces SynthID-Text, a text watermarking system for AI-generated content.

Researchers at Google DeepMind have introduced SynthID-Text, a watermarking system for text generated by large language models like Google Gemini and ChatGPT. This method, which subtly alters word choices, aids in identifying AI-generated content without compromising text quality. While effective, the watermark can be diminished if users edit the text. The technology aims to address concerns in academia and misinformation, though its future integration into AI tools remains undecided.

October 23, 2024
24 Articles