What to know

  • Google announced SynthID, a new tool that can detect AI-generated content, at its I/O 2025 developer conference.
  • SynthID uses invisible watermarking technology that can be embedded in AI-generated images, text, audio, and video.
  • The tool will be available to both creators and consumers through Google's suite of products and as an API for developers.

Google has introduced a new tool called SynthID that can detect AI-generated content. The company unveiled this technology during its annual I/O developer conference on May 21, 2025.

SynthID works by embedding an invisible watermark into content created by AI systems. This watermark doesn't change how the content looks or sounds but allows the tool to identify whether something was made by artificial intelligence.

The technology works across different types of media including images, text, audio, and video. Google says SynthID can maintain its detection capabilities even if the content is modified, cropped, or compressed.

"We want to help people know what's real and what's synthetic," said Google's AI chief during the presentation. "SynthID gives creators and consumers the tools to make informed decisions about the content they encounter online."

Google plans to integrate SynthID into its own AI products first. This means content generated through Google's AI tools will automatically include these invisible watermarks. The company will also make SynthID available as an API for other developers and platforms to use.

The tool comes at a time when distinguishing between human-created and AI-generated content is becoming increasingly difficult. Recent studies show that most people cannot reliably identify AI-created images or text without technological assistance.

SynthID isn't Google's first attempt at content authentication. The company has previously worked on similar technologies, but SynthID represents their most comprehensive approach to date.

Some experts have raised questions about how effective such detection tools can be in the long run. AI systems continue to evolve, potentially making detection more challenging over time.

Google acknowledges these limitations but says SynthID represents an important step toward transparency in AI-generated content. The company plans to continue improving the technology based on user feedback and changing AI capabilities.

The first version of SynthID will be available to select partners next month, with wider availability planned for later this year.