Google expands its tools for verifying AI content

Google is expanding its tools so you can find out where an image, video or audio file came from and whether it was modified with artificial intelligence. Verification is coming to Search and Chrome, while Gemini, Pixel and Google Cloud are also getting stronger capabilities.

The foundation of this strategy is SynthID, a digital watermark that is invisible to the naked eye and that Google embeds in content created with its AI models. According to the company, it has already been applied to more than 100 billion images and videos and 60,000 years of audio.

What you’ll be able to check

Gemini already lets you check whether an image, video or audio file contains a SynthID watermark. Google says the feature has been used 50 million times worldwide, and it will now also come to Search.

It will be added to Chrome in the coming weeks. You’ll be able to use features such as Lens, AI Mode or Circle to Search and ask directly whether content was created with AI. Gemini integrated into Chrome will also be able to help you check where it came from.

The tool does not work as a universal test for every file on the internet. It identifies compatible watermarks, especially those added through tools that use SynthID.

Google is also adding C2PA Content Credentials, a standard that records how a file was created and modified. For example, it can indicate whether a photo came directly from a camera or was later processed with an editor or an AI generator.

The company considers identifying authentic, unedited content just as important as detecting content created or changed with AI.

More information from Pixel phones

The Pixel 10 was Google’s first phone to add these credentials directly from the camera app. In the coming weeks, the feature will expand to video recorded with the Pixel 8, 9 and 10.

This will make it possible to document that an image or video was captured by a camera at the time of recording. If you later share it on Instagram, Meta plans to recognize and display that information using its own provenance labels.

For you, the change could be useful when you receive a viral image, an advertisement or a video whose authenticity is unclear. Instead of relying only on how it looks, you’ll have more clues about its history. Even so, the absence of a credential does not by itself prove that content is fake: many files lose this data when they are copied, edited or posted on platforms that do not preserve it.

A tool for businesses

Google will also launch an AI-generated content detection API on Google Cloud. An API is a connection that lets other programs use a function without building it from scratch.

Companies will be able to use it to analyze content created by Google’s models and other popular models. Possible uses include:

• Sorting posts on a platform.
• Detecting synthetic material in insurance verification processes.
• Helping with fact-checking or labeling AI-generated content.

The service will launch with a group of partners and be adjusted based on their feedback. Google is also working with companies such as OpenAI, Kakao and ElevenLabs so that SynthID can be applied to more AI-generated content.

The direction is clear: content provenance is starting to become an information layer that travels with the file, rather than a decision based solely on how the content looks. The next challenge will be making sure these credentials survive as files move between apps, social networks and editors, and that platforms display them in an understandable way when they really matter.