Google has made its improved SynthID detector available worldwide. The service lets anyone log in with a Google, OpenAI, or Apple account and run a quick check on an image, video or audio file to see if it carries a watermark from those companies' AI generators. Each account gets roughly ten checks per day, although the exact limit can tighten if many similar files are submitted. Google says the restriction helps stop people from reverse‑engineering the detector to create removal tools.
What the tool can and cannot spot
SynthID looks for watermarks that the supported companies embed in AI‑generated media. It does not detect content from other providers that use different watermarking methods, such as Meta, or from open‑weight models that add no watermark at all. The website carries a disclaimer warning users that many AI creations will go unnoticed. Google notes that even with its system, clever attackers may eventually strip the watermark, making the content appear genuine.
Why it matters for everyday users
For anyone trying to verify whether a picture, clip, or sound clip was made by an AI, SynthID offers a single, easy‑to‑use place to check. The daily quota means casual users can still get a quick sanity check without needing special software. However, the tool’s limits and blind spots mean it shouldn’t be the only method relied on for authenticity, especially for content from less‑known sources.
Why it matters
Regular people can now verify the origin of a handful of AI‑generated media each day without installing extra software. The limited quota and focus on only three providers mean the tool is a helpful first step, but users should remain cautious with content from other AI services.
