Google launches SynthID website to check if media was made by AI

In short: Google has launched a public website called SynthID that helps people check whether an image, video, or audio clip was made using AI.

What happened

Google launched a new site at https://synthid.com/ that lets anyone upload media and see if it includes a SynthID watermark. A watermark is like a hidden stamp that is added at the time something is created, similar to an invisible signature.

Google previously offered this checker to a limited group, including journalists, media professionals, and researchers. Now it is open to everyone.

The site supports many common file types. For images, that includes JPG and PNG, plus formats like GIF and TIFF. For video, it supports MP4, MOV, and WEBM, and for audio it supports files like MP3 and WAV.

SynthID is the system Google introduced in 2023 for watermarking AI-made media. Google said its tools and AI models, including Gemini and other media makers it names, add this watermark to what they generate. Google also said verification is built into the Gemini app and the Chrome browser, and that people already make about 1 million verification requests per day.

Other companies also support SynthID, including OpenAI, Nvidia, and Kakao. OpenAI also runs its own verification page. TechCrunch reports Apple is expected to add support soon.

Why it matters

AI-made photos, videos, and audio can spread quickly, especially on social media. A tool like this can help people double-check what they are seeing or hearing. Still, watermark checks are not perfect, and other systems from Microsoft and Meta have sometimes missed AI content, even when it came from their own tools.

Source: TechCrunch AI