Google opens SynthID detector to everyone for AI-generated content verification

Google has rolled out SynthID.com, a public-facing website that allows anyone to upload images, videos, or audio files and check whether they were created with artificial intelligence. The launch marks the company's expansion of a tool that was previously limited to journalists, media professionals, and researchers during its testing phase earlier this year.

How it works#

The detector accepts a broad range of file formats across multiple media types. Users can upload images in JPG, JPEG, PNG, BMP, WEBP, AVIF, HEIC, HEIF, TIFF, or GIF formats; videos as MP4, MOV, or WEBM files; and audio in WAV, MP3, OGG, FLAC, AAC, or M4A. The system relies on SynthID, a watermarking and detection technology that Google introduced in 2023 to identify machine-generated content.

Google's own generative models—including Nano Banana, Veo, and Lyria—embed SynthID watermarks into their output. The same technology powers detection across products with generation capabilities like Gemini, Flow, ProducerAI, and Vids.

Growing industry adoption#

SynthID has gained traction beyond Google's ecosystem. OpenAI, NVIDIA, and Kakao have adopted the standard, while OpenAI maintains its own verification tool. Apple is expected to add support in the coming months.

Google has also integrated SynthID verification directly into the Gemini app and Google Chrome, with the company reporting that users now make roughly 1 million content verification requests daily.

Current limitations#

The technology faces real-world constraints. Microsoft and Meta have developed their own watermarking and verification approaches that operate independently of the SynthID standard. These competing systems, however, have proven unreliable—Meta's detector, for instance, has struggled to identify AI images generated by its own models, particularly when files have been modified.