Google Expands AI in Photos with Video Remix Feature as…

On July 9, 2026, Google's SynthID watermarking system was instrumental in identifying and debunking a widely circulated AI-generated image that falsely depicted Kentucky Senator Mitch McConnell in a hospital bed. The image, which emerged during a period of public speculation regarding the Senator's health, was confirmed to be synthetic after fact-checking organization Snopes identified the presence of a SynthID watermark.

What happened

The incident involved a fabricated image showing Senator Mitch McConnell in a medical setting, which gained traction online amidst ongoing health rumors. The image was identified as AI-generated because it carried the SynthID watermark, a digital signature embedded directly into media created by participating artificial intelligence models. This specific watermark is designed to remain detectable even when content is shared, captured, or reposted across various digital platforms, allowing for verification of its synthetic origin.

Google has been embedding SynthID into images generated by its Gemini models since 2025. The effectiveness of the system in this case highlighted the role of provenance technology in combating misinformation. While OpenAI joined the SynthID program in May 2026 to help flag AI-generated content, other entities, such as Anthropic, do not currently participate in the system. The successful identification of the McConnell image serves as a practical demonstration of how persistent watermarking can provide transparency for digital media that has been manipulated or generated by AI tools.