AI news videos

Google Opens SynthID Detector Access to Everyone in English

Google DeepMind published a post on blog.google on October 7, 2026, by Pushmeet Kohli, VP, Science and Strategic Initiatives, saying the SynthID Detector is now available to everyone, globally, in English starting today.

Watch on YouTube

Transcript

Sources

Every claim in this video comes from the top ranking coverage of this topic. The claims and where each one came from:

• Backing sentences, quoted from the post. HOOK BACKING, that access is what is new and the tool is not: "Last year, we introduced an early version of the SynthID Detector to help media professionals verify AI-generated content. Now, we're expanding access to everyone, with the tool available globally in English starting today." The hook's claim that the tool is not new is further backed by "Since launching Google SynthID in 2023, we've helped people identify AI-generated media using imperceptible watermarks across images, video, and audio." SCENE 1 (an early version went to media professionals last year, and the detector is now available globally in English): same two sentences as the hook backing, "Last year, we introduced an early version of the SynthID Detector to help media professionals verify AI-generated content. Now, we're expanding access to everyone, with the tool available globally in English starting today." SCENE 2 (anyone can check whether a piece of media was made with AI from Google or a partner): "With the new SynthID Detector, anyone can easily check if an image, video, or audio file was made with AI from Google or our partners, including OpenAI, NVIDIA, Kakao, and soon, Apple." SCENE 3 (SynthID is not new, and watermarking already covers more than one hundred eighty billion images and videos): "Since launching Google SynthID in 2023, we've helped people identify AI-generated media using imperceptible watermarks across images, video, and audio. To date, we've watermarked over 180 billion images and videos, along with 240,000 years of audio content." SCENE 4 (no accuracy rate, English only, no text coverage): backed by the verified absences in the capture, plus the published scope sentence "Now, we're expanding access to everyone, with the tool available globally in English starting today" and the media-type list "an image, video, or audio file". SCENE 5 (Google products already handle more than one million checks every day, and the figures are Google's own): "This tool joins our built-in verification features in Search, the Gemini app, and Chrome, which now regularly handle over 1 million requests daily." Closing context sentence on the page: "It's all part of our ongoing work to give you more context about the media you see online, so you can navigate the web with confidence." WHAT IS NOT IN THE SOURCE, stated explicitly: there is no accuracy, false-positive or false-negative rate for the detector; there is no price and no mention of whether an account or sign-in is required; there is no tool URL published in the body, only anchor text linking to the 2023 SynthID launch, to last year's early Detector announcement and to the tool; there is no date for any language beyond English; there is no date for the promised Apple support, which the post describes only as "and soon, Apple"; there is no list of which OpenAI, NVIDIA or Kakao models are covered; there is no claim of detection support for AI-generated TEXT, the listed media types being images, video and audio only; and there is no independent or third-party verification of the 180 billion images and videos, the 240,000 years of audio content, or the 1 million daily requests, all three of which are Google's own numbers. NOTE ON QUOTING: the source sets off the clause "to everyone - with the tool available globally in English starting today" with an en-dash, and a QA gate bars every dash form from every field of this file, so that en-dash is rendered here and in onscreen text as a comma; hyphenated source compounds such as "AI-generated" and "built-in" are preserved only inside quoted onscreen and source text and never appear in narration, where the phrasing is "AI generated" and "built into". Curly apostrophes are rendered as straight ones. Wording is otherwise verbatim. NOTE ON NUMBERS: narration speaks exactly two figures, each once and in non-adjacent scenes, "more than one hundred eighty billion images and videos" in scene 3 and "more than one million checks every day" in scene 5, both written as words rather than digits because digits break our transcription audit; the 240,000 years of audio content figure and the 2023 SynthID launch year are carried in onscreen text only and are never narrated, because two spoken figures close together send our audio transcription audit into a repetition loop. NOTE ON ENUMERATIONS: the partner list (OpenAI, NVIDIA, Kakao, and soon Apple), the media-type list (images, video, audio) and the product list (Search, the Gemini app, Chrome) are carried in onscreen text only, because multi-item lists are spoken far more slowly than prose and overrun the 60 second limit; "Kakao" is additionally absent from our pronunciation vocabulary and is therefore never narrated.(HOOK AND FULL-SCRIPT BACKING. First-party Google post on blog.google, in Innovation & AI > Models & research > Google DeepMind, headline "We're making it easier to identify AI-generated content globally.", written by Pushmeet Kohli, VP, Science and Strategic Initiatives, Google DeepMind, and dated Oct 07, 2026 on the page. Posted in the Google DeepMind and Safety & Security categories. Our watcher captured a snapshot of the page at 2026-10-07T14:01:11.169Z, and the live URL was verified the same day, 2026-10-07, with the live text matching the alert-time snapshot verbatim. The captured body is 180 words and the remainder of the page is site navigation chrome only. Zero corroborating YouTube transcripts were captured for this story (0 of 5), because yt-dlp is out of date on this host, which is a known tooling issue and not a content problem, so this single first-party post is the sole fact source for every narrated claim in this Short, and no secondary reporting, analyst note, benchmark or third-party test was fetched or checked. No prompt-injection or instruction-like content was found in the fetched article, the snapshot or the ranking-research file.)