As deepfakes become increasingly harder to recognize, Google has launched SynthID, a cutting-edge tool designed to help individuals identify AI-generated images and curb misinformation.
In a blog post published on Tuesday, August 29, Google DeepMind said SynthID can spot images created with Imagen, Google’s text-to-image AI model.
“This technology embeds a digital watermark directly into the pixels of an image, making it imperceptible to the human eye, but detectable for identification,” Google DeepMind said.
This development comes amid growing concerns over the malicious use of deepfake images generated by AI.
“While generative AI can unlock huge creative potential, it also presents new risks, like enabling creators to spread false information — both intentionally or unintentionally. Being able to identify AI-generated content is critical to empowering people with knowledge of when they’re interacting with generated media, and for helping prevent the spread of misinformation,” Google DeepMind explained.
The beta version of SynthID is currently available to “a limited number of Vertex ID customers using Imagen.”
Identifying AI-Generated Content
SynthID relies on two deep-learning models to accurately identify watermarked images and enhance the watermark’s invisibility by visually aligning it with the image.
According to Google DeepMind, traditional watermarks often fall short when it comes to AI-generated images, as the watermarks can be easily edited out and are not designed to survive image manipulations.
SynthID takes watermarking to the next level. It doesn’t compromise image quality, and the watermark remains detectable even after filters are applied to the image.
While the “technology isn’t perfect,” Google DeepMind said it can accurately detect AI-generated images even after they’ve been edited.
“SynthID isn’t foolproof against extreme image manipulations, but it does provide a promising technical approach for empowering people and organisations to work with AI-generated content responsibly,” Google’s blog post reads. “This tool could also evolve alongside other AI models and modalities beyond imagery such as audio, video, and text.”
Ensuring Responsible Use of AI-Generated Content
SynthID offers a powerful and scalable method for mitigating risks associated with deepfakes. The technology is adaptable and could evolve to work with other forms of AI-generated content like audio, video, and text. As such, it could also help to curb rising AI voice scams.
DeepMind is collecting feedback to enhance the tool and is considering integrating it into more Google products and third-party solutions in the near future.
The launch of SynthID represents a significant advancement in combating misinformation and fulfills recent commitments by leading AI companies, including Google and OpenAI, to adopt responsible AI practices. It also aligns with the EU’s upcoming AI Act.
Watch the video below for more information about SynthID:
For more cybersecurity news, follow us on X (Twitter), Threads, and Mastodon!
