Google’s SynthID system has been used to debunk a high-profile AI-generated fake image, in a rare but significant victory for the system.

Earlier this week, an image circulated online that appeared to show Kentucky Sen. Mitch McConnell covered in tubes in a hospital bed in a state of extreme distress. The image was widely shared on Reddit and

In short, the watermark worked exactly as it was supposed to in a victory for anti-deepfake technology.

Senator McConnell’s health has been the subject of intense speculation since he was admitted to the hospital following an emergency call on June 14. Since then, he has been largely absent from the public view, fueling speculation that his health could be worsening. In this case, however, the evidence turned out to be completely false.

Launched at Google’s I/O developer conference in 2025, SynthID works as an invisible signature, visible to SynthID algorithms but designed to go unnoticed by the casual observer. Because the signature is embedded in the image itself, it survives even when an image is captured on multiple platforms, as was the case with McConnell.

The main limitation of SynthID is that it can only be used when an imaging tool is actively involved in the program. Gemini models have included the watermark since the program launched in 2025. OpenAI joined in May 2026, as part of a broader effort to combat malicious image generation. Anthropic does not participate in the program.

Users can check whether images contain the watermark by asking a Gemini model or by uploading them to OpenAI’s public image verification tool.