Google unveils plans to help you tell real images from fakes

Google plans to introduce technology that will recognize whether a photo was taken with a camera, edited with software like Photoshop, or generated by generative AI models. In the coming months, Google search results will include an updated “about this image” feature so users know if an image was created or edited with AI tools.

The system Google uses is part of the Coalition for Content Provenance and Authenticity (C2PA), one of the largest groups trying to address AI-generated images. C2PA is a technical standard that includes information about the provenance of images and works in both hardware and software to create a digital footprint. Amazon, Microsoft, Adobe, Arm, OpenAI, Intel, Truepic and Google have all endorsed C2PA, but adoption has been slow. Google’s integration with search results will be the first big test for the initiative.

Google helped develop the latest C2PA technical standard (version 2.1) and will use it alongside the upcoming C2PA Trust List, which allows platforms like Google Search to confirm the origin of content. “For example, if the data shows that a photo was taken by a specific camera model, the Trust List helps confirm that this information is accurate,” says Laurie Richardson, vice president of Trust and Security at Google.

Google also plans to integrate C2PA metadata into its advertising systems. “Our goal is to gradually scale this up and use C2PA signals to inform how we enforce key policies,” Richardson says. “We’re also considering ways to provide C2PA information to YouTube viewers when content is captured on camera, and we’ll have more on that later this year.”

While Google stands out as one of the first major tech companies to adopt the C2PA authentication standard, there are still many challenges ahead in terms of adoption and interoperability to make the standard work across a wide range of hardware and software. Only a handful of cameras from Leica and Sony support the open C2PA technical standard, which adds metadata about camera settings, as well as data and location of the shot, to photos. Nikon and Canon have committed to adopting the C2PA standard, and we’re still waiting to hear whether Apple and Google will implement C2PA support on iPhones and Android devices.

Adobe Photoshop and Lightroom can add C2PA data, but Affinity Photo, Gimp, and many others can’t. There are also challenges with how to display the data once it’s added to a photo, since most major online platforms don’t offer labels. However, Google’s adoption in search results could encourage others to implement similar labels.

“Determining and signaling content provenance remains a complex challenge, with a range of considerations based on the product or service,” Richardson acknowledges. “And while we know there is no silver bullet solution for all online content, working with others in the industry is key to creating sustainable and interoperable solutions.”