Google's Circle to Search can now detect AI-generated images
What's the story
Google is expanding its artificial intelligence (AI) detection capabilities with the update of SynthID, its proprietary watermarking system. The invisible metadata tagging system is used for content created or modified with Google's own AI tools. The tech was first introduced at I/O last year and later integrated into the Gemini app. Now, Google is bringing these capabilities to Chrome and Search as well.
Detection process
How to check if an image is AI-generated
With the latest update, users can use SynthID to check an image's origin in multiple ways. On Android, Google's Circle to Search will flag AI content, while Google Lens and Chrome's version of Gemini will answer questions like "is this AI generated?" The tools will provide granular details about a given image. For instance, Google explained that an image was originally taken on a Pixel phone but later edited with AI-enabled tools in the Google Photos app.
Scope and reliability
Limitations of Google's detection capabilities
Google can't guarantee the same level of detail for every potentially AI-altered image. The company has added content credentials, an industry-standard watermarking system, to the native camera app in its Pixel 10 lineup. This tech is now being expanded to Pixel 8 and 9 models. While Google can trace images captured and edited via its own tools, it may not be as reliable with content from other AI platforms.
Partnerships
Deals with OpenAI, Kakao, and ElevenLabs
Google has said that OpenAI, Kakao, and ElevenLabs have decided to bring "SynthID technology to more of their AI-generated content." OpenAI said its integration would start with pictures created using "ChatGPT, Codex, or the OpenAI API." Google is also expanding support for content credentials in a bid to make AI detection more universal. The Gemini app will get content credential support starting today, while Chrome and Search are expected to follow suit in the coming months.