Google’s deepfake detector system used to debunk McConnell hoax pic

Google’s SynthID The system has been used to debunk a high-profile AI-generated hoax image, marking a rare but important win for the system.

Earlier this week, a photo circulated online showing Kentucky Senator Mitch McConnell covered in tubes appearing to lie in a hospital bed in extreme distress. The image was shared widely on Reddit And Xbut by Wednesday, the respected fact-checking site Candy had debunked the image, noting that when checked, the image records as containing the SynthID watermark designed by Google to identify AI-generated images.

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

Senator McConnell’s health has been the subject of intense speculation since he checked into the hospital after an emergency call on June 14. Since that time, he has been largely absent from the public eye, fueling speculation that his health may be declining. However, in this case the evidence turned out to be completely false.

Launched at Google’s 2025 I/O developer conference, SynthID acts as an invisible signature, visible to SynthID algorithms but designed to be unnoticeable to the casual observer. Because the signature is built into the image itself, it remains intact even if an image is captured on multiple platforms, as with McConnell’s image.

The main limitation of SynthID is that it can only be used if an image generation tool is actively participating in the program. Gemini models have included the watermark since the program’s launch 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 if images contain the watermark by asking a Gemini model or by uploading them OpenAI’s public image verification tool.