The biggest names in artificial intelligence (AI) have rallied around an Adobe and Microsoft-led initiative to standardise content metadata tagging in the hopes of making it easier to spot AI-generated content online.

Google, Meta, and OpenAI each announced efforts in the past week that further entrench the Coalition for Content Provenance and Authenticity (C2PA) as the leading standards body for attaching invisible data to digital media.

C2PA is in charge of the 'Content Credentials’ standards which is a method for embedding images, video, and audio with provenance metadata – information like who created it, how, with what technology, and whether the content has been edited.

That data is cryptographically signed in a way that C2PA says makes it “tamper-resistant”.

On Thursday, Google announced that it was joining the C2PA’s steering committee alongside Adobe, BBC, Microsoft, global marketing outfit Publicis Groupe, Sony, and Truepic.

Google VP of Trust and Safety, Laurie Richardson, said Content Credentials will fit into the company’s toolbox of ways to try and spot synthetic material on YouTube while making sure its generative AI products carry the appropriate metadata.

“At Google, a critical part of our responsible approach to AI involves working with others in the industry to help increase transparency around digital content,” she said.

“This is why we are excited to join the committee and incorporate the latest version of the C2PA standard.”

Google’s announcement came just days after OpenAI – creator of ChatGPT – said it would put C2PA’s standard into its DALL-E 3 image generator.

OpenAI said C2PA’s metadata approach was “not a silver bullet to address issues of provenance,” however, noting that C2PA “can be easily removed”.

“For example, most social media platforms today remove metadata from uploaded images, and actions like taking a screenshot can also remove it,” OpenAI said.

“Therefore, an image lacking this metadata may or may not have been generated with ChatGPT or our AI.”

Open AI said the inclusion of C2PA would have a “negligible” effect on latency to its products, turning a 3.1MB PNG file created through its API into a 3.2MB image.

Earlier in the week, Meta said it would integrate C2PA-developed techniques into Facebook, Instagram, and Threads to let users see when content in their feeds was created by AI.

It’s a step toward stemming the tide of AI-generated noise that risks further flooding social feeds with disinformation and scammers.