TechCrunch reported in a post on X that Google has launched a website allowing people to check whether an image, video or audio clip was made with artificial intelligence. The announcement arrives as fabricated recordings increasingly move beyond experimental demonstrations and into scams, impersonation attempts, political messaging and misleading online posts.
The tool is called SynthID Detector. In Google’s announcement, the company said the portal scans uploaded content for invisible SynthID watermarks. It can also identify areas of a file that are likely to contain the watermark, giving users more information than a simple yes or no result.
A watermark rather than a universal truth
SynthID works by embedding an imperceptible signal into content produced by Google’s AI systems. According to Google DeepMind’s explanation of SynthID, the technology can watermark AI-generated images, audio and video while keeping the signal difficult for people to see or hear during normal use.
That distinction matters. The detector is not presented as a universal test for every piece of synthetic media on the internet. It is designed to find SynthID, which means its strongest results will come from content created or edited with Google tools that use the watermarking system.
A file made by another company’s model may not contain the signal. The same may be true if an image, recording or video has been heavily altered, compressed or re-edited. In those cases, a negative result would not necessarily prove that the material is authentic. It could simply mean that Google’s watermark is absent or no longer detectable.
Google initially opened access to early testers, including journalists, media professionals and researchers, according to its announcement. Their work reflects the groups most likely to encounter questionable material before it reaches a wider audience. Reporters may use the service when examining a suspicious recording, while researchers and publishers could incorporate it into broader systems for checking digital provenance.
A useful signal for people under pressure
For ordinary users, the appeal is straightforward. A video that appears to show a public figure making an inflammatory statement, or an audio clip that seems to capture a private conversation, can spread faster than anyone can independently verify it. A quick check could encourage people to pause before sharing or acting on the material.
Google’s separate Gemini help page explains that signed-in users can upload an image, video or audio file and ask Gemini to check it for SynthID. A positive result indicates that all or part of the content was created or edited by Google AI. That wording is important because a genuine photograph, recording or video may have been modified with an AI feature without being generated entirely from nothing.
The detector therefore works best as one piece of evidence, not as a replacement for reporting, source verification or human judgment. Users still need to ask where a file came from, whether the account sharing it is credible and whether independent evidence supports its claims.
The larger contest over trust
Google’s move also highlights a wider struggle over responsibility in the generative AI industry. As synthetic media becomes cheaper and more persuasive, platforms and model developers face growing pressure to make the origins of digital content easier to understand.
Watermarking offers one possible answer, but its value depends on adoption, durability and transparency. If companies use different systems, or if their signals disappear during routine editing, no single detector will settle every dispute.
Still, a public verification service gives journalists, researchers and the public a practical starting point. In an online environment where seeing is no longer enough to believe, even an imperfect signal may help people slow down and look twice.
- Austin McKinley · CC BY 3.0
Daniel Reyes is not a person. No notebook, no deadlines, no face behind the name — just a byline this newsroom publishes under. Here is the production line underneath it, because a name beside a portrait reads like a journalist, and this one is not one.
The models. Writing: gpt-5.6-luna and qwen3-max. Out on the live web: gpt-5.6-luna and gpt-5.6-terra. Pictures: gpt-image-1 and gpt-image-1-mini. Swap one in the newsroom and this line swaps with it — it is read off the machines, not typed here.
How a story is made
- Research. The searching model reads around the story, pointed at primary sources — the filing, the post, the repository — rather than at somebody else's write-up of them.
- Writing. The writing model drafts it against what was found, at Daniel Reyes's usual length and in Daniel Reyes's usual register.
- The loop. A reviewer reads the draft and sends it back with notes. Then reads it again. A piece can go round several times before it leaves the building.
- Enrichment. A quotation has to appear word for word on the page it is taken from. A chart may only use figures that appear in the source it cites. Whatever fails is dropped, and the reason is kept.
- Fact check. A last pass hunts for claims the article makes and its sources do not.
- A human stop. Sensitive subjects are held for a person to read before publication, and a person can kill any of it at any point.
If that sounds less like a newsroom and more like a factory: quite. It is called Press Factory.
This article was generated using AI and published automatically without human pre-publication review.
How this article was made
The article was produced by the Grandmonts Media News Engine using automated research, drafting and verification workflows. No human editor reviewed the article before publication. Grandmonts Media remains responsible for the published content. Errors can be reported at office@grandmonts.cz.