Google adds new schema to label AI-generated content

Google has added a way to label whether content is AI-generated via a new schema property - digitalSourceType. This has been added to the docs pages for Q&A and Forums. With it, you can now tell big G if the content on the page was generated by an LLM (TrainedAlgorithmicMediaDigitalSource) or a simpler algo (AlgorithmicMediaDigitalSource). Not specifying a value means that they will apparently assume the content is written by humans.

The question is, who the eff is going to be specifically labelling their content as AI-generated? And how long until they add it to all content types?

Interesting update, but the question is fair In reality, most people probably won’t voluntarily label their content as AI-generated — for now, it’s more about transparency for large platforms and brands. I think over time algorithms may start taking this into account, and then it won’t be so easy to ignore.

Most website owners are unlikely to label their content as AI generated right away,dice it may affect SEO and how user perceive their site.However,if google begin digitalsourcetype as ranking.

Nobody on BHW is going to voluntarily label their content as AI-generated, that's the whole point. This schema is designed for publishers who want to be transparent - news orgs, academic sites, places that actually care about disclosure. The rest of us will just not add it and Google assumes human-written by default, which is exactly where you want to be.

The more interesting question is what happens when Google starts detecting it automatically rather than relying on self-reporting. Because that's obviously where this is going. The schema is just the first step to train their own detection systems - once they have enough labeled examples of confirmed AI content they'll use that to identify unlabeled AI content.

So the real move is not worrying about the schema itself, but making sure your AI content doesn't read like AI content. Which is a whole different problem.

They are onto something again.

Looks like a foreplay of a related penalty or filter. (more strict than the current ones)

The schema is more about structuring trust signals than forcing compliance. Google explicitly says if you don’t label it, they’ll just assume it’s human
Right now it’s only for forums/Q&A, but it’s clearly the first step toward broader rollout
Real play here is future ranking and AI citations, not whether people honestly tag their content.