Google opens SynthID website for checking AI-generated media

Google has launched a new SynthID website that lets anyone check whether an image, video or audio clip may have been generated with AI, according to TechCrunch. The broader rollout matters for creators, marketers and small businesses that need a practical way to assess media authenticity before publishing, licensing or sharing content.

The company previously made the tool available to selected journalists, media professionals and researchers for testing around Google I/O last year. Now, Google is opening access beyond that limited group.

What the new SynthID site checks

The website uses SynthID, Google’s watermarking and detection system introduced in 2023, to look for signals in AI-generated media. Google applies SynthID watermarks to media produced by its own generation models and tools, according to the report.

For readers working with mixed media assets, the most important detail is format support. The site can check common image, video and audio files, including:

• Images: JPG, JPEG, PNG, BMP, WEBP, AVIF, HEIC, HEIF, TIFF, TIF and GIF
• Video: MP4, MOV and WEBM
• Audio: WAV, MP3, OGG, FLAC, AAC and M4A

That makes the site relevant to a wide range of everyday publishing workflows, from product images and campaign videos to podcast clips and social posts. It does not mean every AI-generated file will be identified, but it gives teams a dedicated place to run checks on supported files.

How it connects to Gemini and Google’s AI tools

Google says SynthID is used by its Nano Banana, Veo and Lyria models, as well as generation tools including Google Gemini, Flow, ProducerAI and Vids. Media created through those systems is watermarked with SynthID, which is what the new site is designed to detect.

Google has also built SynthID verification into the Gemini app and Google Chrome. According to the company, people currently make 1 million content verification requests each day.

For creators already using Gemini or other Google AI products, the standalone site adds a more direct verification option. For agencies and small businesses, it may be useful as a checkpoint when reviewing supplied media, especially when files come from freelancers, clients, stock sources or internal AI workflows.

Why media verification is becoming a workflow issue

The launch lands in a market where AI-generated media is increasingly easy to create and harder to assess by sight or sound alone. A public verification site does not replace editorial judgment, rights checks or source verification, but it can become one layer in a content review process.

For marketers, that could mean checking a visual before it is used in an ad campaign. For publishers and newsletters, it could mean screening images or audio clips before publication. For freelancers, it could help document whether a submitted file appears to carry a SynthID signal.

The key limitation is scope. SynthID verification is most directly tied to media watermarked with SynthID. TechCrunch notes that watermarking and verification tools are not foolproof, and can fail to identify content created by the same companies’ models. That is an important caution for teams that might be tempted to treat any single checker as a final answer.

A fragmented verification landscape

Google is not alone in building AI media verification systems. OpenAI, Nvidia and Kakao support SynthID, according to the report, while OpenAI also operates its own content-checking site. Apple is reportedly expected to add support as well.

At the same time, Microsoft and Meta use their own standards for watermarking and verifying content. That means the field remains fragmented: a file may need to be checked against different systems depending on how it was created, edited or distributed.

For Emonarc readers, the practical takeaway is to treat Google’s new site as a useful addition, not a universal detector. If your business publishes AI-assisted media, the safest workflow is still to keep clear records of how assets were made, label AI-generated content when appropriate, and use verification tools as supporting evidence rather than the sole source of truth.