Google Now Wants to Know If Your Forum Content Was Written by AI — An SEO Signal

On March 24, 2026, Google updated its official documentation for Discussion Forum and Q&A Page structured data, according to Search Engine Journal. The key addition: a digitalSourceType property that allows forum and Q&A sites to declare when content was generated by AI. It's quiet. It could be very significant.

What Google Actually Changed

The new digitalSourceType property uses IPTC digital source enumeration values — the same system Google already applies to images. For forum and Q&A content, it supports two values:

• TrainedAlgorithmicMediaDigitalSource — content generated by a trained AI model (GPT, Gemini, Claude, etc.)
• AlgorithmicMediaDigitalSource — content produced by a simpler algorithmic system (auto-reply bots, templates)

The property is recommended, not required. If you don't declare it, Google defaults to assuming the content is human-generated. Which is precisely where it gets interesting.

Google also added commentCount as a recommended property (total comment count, even when paginated) and expanded the types supported by sharedContent to include referenced posts from other threads.

Why Google Is Doing This Now

Forums like Reddit, Quora, and product Q&A sections are heavily used as sources for Google's AI Overviews. Google needs to know what it's ingesting — is this a genuine community discussion, or machine-generated output?

The stakes are UGC signal authenticity. A forum filled with AI-generated answers doesn't carry the same value as one where 200 users have been debating for five years. Google hasn't disclosed how it will use digitalSourceType in ranking or display — but extending the concept from images to text is a strong directional signal.

What This Means for Your Content Strategy

For businesses using chatbots or LLMs to answer FAQs on their websites, this update raises a real strategic question — and it's directly relevant to how you approach SEO copywriting in the AI era. It's a critical inflection point for any SEO content strategy in 2026: where does automation end and human expertise begin? Two approaches:

• Declare AI content via digitalSourceType — transparent and compliant, but you're explicitly signaling to Google that your answers are automated
• Ensure human oversight — an expert reviews and validates each AI-generated response before publication, justifying the omission of the label (the human value-add is real)

For E-E-A-T signals, the second approach is clearly preferable. Google rewards human experience and expertise. An entirely AI-generated FAQ — even well-written — doesn't carry the same proof of expertise as an answer written by someone who actually solved the problem.

What to Do Now

If you have a forum, Q&A section, or user comments on your site, here are the concrete steps:

• Audit your UGC: what proportion is AI-generated or AI-assisted?
• If your bot automatically answers questions → add digitalSourceType to your markup
• If a human reviews and validates responses → don't label, but document the process (useful if Google ever audits your E-E-A-T)
• Add commentCount if you have paginated comment sections — it's an easy engagement signal to implement

Cicero's take: Google is quietly building infrastructure to distinguish human content from AI content at scale. This structured data update isn't trivial — it's one link in a longer chain. Sites that play it transparently now will be in a better position when the lever is activated in ranking. This aligns directly with the E-E-A-T criteria Google keeps reinforcing with each algorithm update.

Alexis Dollé

CEO & FOUNDER

Growth and SEO content strategy specialist, I founded Cicéro to help businesses build lasting organic visibility — on Google and in AI-generated answers. Every piece of content we produce is designed to convert, not just to exist.

LinkedIn