Google Creates a Watermark for AI Images That Can’t Be Removed (1)
Google DeepMind, led by CEO Demis Hassabis, has unveiled SynthID, a tool designed to watermark AI-generated images in a manner undetectable to the human eye but easily identifiable by AI detection tools. This watermark remains intact even after the image undergoes transformations like cropping or resizing.
An industry wide standard would be interesting. But getting private enterprise to enforce that standard is another thing, unless federally regulated. We will file this under “the world’s ongoing effort to regulate AI and creating separation between human and AI generated works”.
The company is initially introducing SynthID to Google Cloud customers using the Vertex AI platform and the Imagen image generator. Hassabis envisions SynthID becoming an internet-wide standard, potentially extending its application to video and text. However, Google isn't alone in this endeavor; other tech giants are also developing AI detection mechanisms. SynthID's debut coincides with Google's Cloud Next conference, where the increasing use of the Vertex AI platform was highlighted. As AI becomes more integrated into daily business, tools like SynthID will play a crucial role in distinguishing genuine content from AI-generated material.
