Google launches SynthID website to check media for AI watermarks
Google has opened a dedicated SynthID website where anyone can check images, video and audio for the AI watermark, now covering partners such as OpenAI, Nvidia and Kakao. Coverage notes that content from tools without SynthID will not be flagged.
1 / 2
The story, neutrally told
Centre · 1Google launched a new website that lets anyone check whether an image, video or audio clip was generated using AI, according to TechCrunch, which dated the launch to Tuesday. TechCrunchN “Google on Tuesday launched a new site that lets anyone verify whether a piece of media, be it an image, a video or an audio clip, is generated using AI.” Read at TechCrunch ↗ Left · 1The site is at SynthID.com and checks for SynthID, an invisible signal encoded in the pixels of AI-generated images and videos and in the waveform of AI audio, according to Ars Technica. Ars TechnicaLC “SynthID is an invisible signal encoded in the pixels of AI-generated images and videos, and in the waveform of AI audio.”“you can go to the new SynthID.com website to check a piece of content for an AI watermark” Read at Ars Technica ↗ Mixed · 2Google had given the tool to select journalists, media professionals and researchers for testing at Google I/O last year; before this launch, people outside that trusted tester programme could only check for SynthID by asking Gemini. TechCrunchN “The company had made this tool available to select journalists, media professionals and researchers for testing at Google I/O last year, and is now opening up access to the tool to everyone.” Read at TechCrunch ↗ Ars TechnicaLC “Until now, the only way for those outside of Google's trusted tester program to check for SynthID was to ask Gemini.” Read at Ars Technica ↗
Mixed · 2The detector now supports the watermarks of Google's SynthID partners, including OpenAI, Nvidia and Kakao, with Apple expected soon, so one site can check media from several AI systems. Previously Google's detector only recognised Google's own version of SynthID. Ars TechnicaLC “Google says that its expanded version of the detector will support the distinct watermarks from all its SynthID partners, including OpenAI, Nvidia, Kakao, and (soon) Apple.” Read at Ars Technica ↗ TechCrunchN “OpenAI, NVIDIA and Kakao also support SynthID, and OpenAI maintains its own site for checking content. Apple is said to be adding support soon, too.” Read at TechCrunch ↗ Centre · 1The site accepts common formats: JPG, PNG, WEBP, HEIC and GIF among others for images; MP4, MOV and WEBM for video; and WAV, MP3, OGG, FLAC, AAC and M4A for audio. TechCrunchN “MP4, MOV, and WEBM for videos; and WAV, MP3, OGG, FLAC, AAC, and M4A for audio.” Read at TechCrunch ↗ Mixed · 2Google's Nano Banana, Veo and Lyria models, and tools such as Gemini, Flow, ProducerAI and Vids, watermark their output with SynthID; Google says Gemini alone has integrated SynthID with over 180 billion images and videos and 240,000 years of audio. TechCrunchN “The company's Nano Banana, Veo and Lyria models, as well as tools with generation capabilities, like Gemini, Flow, ProducerAI and Vids, all use SynthID to watermark any media they generate.” Read at TechCrunch ↗ Ars TechnicaLC “Google says Gemini alone has integrated SynthID with over 180 billion images and videos, plus 240,000 years of audio.” Read at Ars Technica ↗
Centre · 1Verification is also built into the Gemini app and Chrome, and Google said people currently make 1 million verification requests a day. TechCrunchN “Google has baked SynthID verification into the Gemini app and Google Chrome. The company said people currently make 1 million requests to verify content every day.” Read at TechCrunch ↗ Left · 1Users must log in with a Google, OpenAI or Apple account, and Google says each gets a daily quota of approximately 10 image, video and audio checks. Ars Technica reports Google engineers said limits are needed to stop people using the checks to develop SynthID removal tools. The site only says whether SynthID was used to produce or edit the media, without the highlighted regions of the internal tool. Ars TechnicaLC “You have to log in to use the tool, but it supports Google, OpenAI, and Apple accounts.”“Google engineers have told Ars that limiting access is necessary to prevent people from using the checks to develop SynthID removal tools.”“Instead, you'll just get a message telling you whether SynthID was used to produce or edit the media.” Read at Ars Technica ↗ Mixed · 2Both outlets note limits: Microsoft and Meta use their own watermarking and verification standards, and TechCrunch says these tools are not infallible. Ars Technica adds that other providers use no watermarks, open-weight models can produce unlabelled content, and someone may find a way to remove SynthID; it says the site will carry a disclaimer, and suggests cryptographic labelling of real media with C2PA, as on Google's Pixel phones, may be a better long-term approach. TechCrunchN “Microsoft and Meta have their own standards for watermarking and verifying content. However, these tools are not infallible, often failing to identify content created by their makers' models.” Read at TechCrunch ↗ Ars TechnicaLC “The best approach might be the opposite—using cryptographic technologies like C2PA to label real images, video, and audio as they are created.” Read at Ars Technica ↗
Every sentence links to the reporting it rests on. The pill in front of each says where its sources sit: Left, Centre or Right when one side supplies at least half of them, Mixed when they are evenly split. The number is how many outlets it cites.
Left1 outlet
- Framing
- Ars Technica treats the site as a useful but imperfect step, explaining how it works and its limits, and ends by calling SynthID 'the best bet you've got' while favouring C2PA labelling of real media.
- Emphasis
- Login requirement, daily quota of about 10 checks, anti-removal rationale, less detail than the internal tool, and gaps from non-SynthID and open-weight models.
- Leaves out or plays down
- Does not mention Chrome integration or the 1 million daily verification requests.
- Charged language
- “rather clunky”“convoluted process”
- For example
-
“Expanding access to SynthID detection is a good thing—it's a convenient single destination to check media for the signature of popular AI tools.” — Ars Technica
“In the meantime, SynthID is the best bet you've got to identify AI fakes.” — Ars Technica
Centre1 outlet
- Framing
- TechCrunch gives a short news report of the launch, leading with public access and listing supported file formats and partners.
- Emphasis
- Supported formats, Google models using SynthID, Chrome and Gemini integration, and 1 million daily requests.
- Leaves out or plays down
- Does not mention the login requirement, usage quota, or that the site gives only a yes/no result.
- For example
-
“Google has baked SynthID verification into the Gemini app and Google Chrome.” — TechCrunch
“The site supports JPG, JPEG, PNG, BMP, WEBP, AVIF, HEIC, HEIF, TIFF, TIF, and GIF formats for images” — TechCrunch
Right0 outlets
No right outlet in our sources has covered this story yet.
What every side reports
- Google launched a SynthID verification website open to the public.
- OpenAI, Nvidia and Kakao support SynthID, with Apple expected to add support.
- SynthID verification already exists in the Gemini app, and Google's tool was previously limited to testers.
- Watermark detection cannot identify content from tools that use other or no watermarking.
Google organisation
Google presents the site as opening SynthID detection to everyone, covering partner watermarks, while rate-limiting use to prevent building removal tools.
“Google has announced that its SynthID detector now supports watermarks from all its partners, and it's available for anyone to use on a new dedicated website.” — Ars Technica
“The company said people currently make 1 million requests to verify content every day.” — TechCrunch
SynthID product
SynthID, introduced in 2023, is a watermark embedded in AI media by Google and partners; coverage says it works only for content made with tools that use it.
“Google is using SynthID, which it introduced in 2023, to identify whether media is AI-generated.” — TechCrunch
“That means there is plenty of AI content that will not be identified as AI in the SynthID tool.” — Ars Technica
Left1 article
-
Google rolls out improved SynthID AI content detector, now available globally
Mixed Explanatory piece welcoming the consolidated detector while detailing its quota, limits and gaps, and suggesting C2PA as a complement.
Centre1 article
-
Google’s new SynthID website can identify AI-generated media
Neutral Brief news report on the launch, focused on access, formats, partners and usage.
Right0 articles
No coverage yet.
- 7 Oct 15:00 First TechCrunchN Google’s new SynthID website can identify AI-generated media
- 7 Oct 15:00 First Ars TechnicaLC Google rolls out improved SynthID AI content detector, now available globally
Times are when each article was published, or when we first saw it if the outlet gave no time.