Google is staffing up to generate and recommend AI-created stories inside Discover. Several listings on Google’s careers site, first shared by Damien Andell on X, point to a feed that will not just rank existing pages but also produce content for what Google calls proactive surfaces.
No timeline or feed mix is spelled out. But the direction is clear enough for search and content teams to start watching now.
What the roles describe
A Staff ML Engineer listing for the Discover Feed Recommendation team says the group “leads by recommending AI-generated content (AIGC) to fulfill user interests.” The role covers topic discovery, content understanding, and quality control, including checks for factuality, freshness, coherence, and tone. It also mentions agentic flows that create content for Discover and Notifications.
A Senior Staff ML role repeats the mandate at a more senior level. The design side is just as telling: a Principal UX listing at Google’s Director+ level, based in Mountain View or New York, pays a U.S. base salary of $275,000 to $382,000 plus bonus and equity. It asks the hire to shape Discover as a “proactive, hyper-personalized concierge experience” that moves people from inspiration to action.
The design brief also says the work will span YouTube, Shopping, Chrome, and Android, and that the hire should scale AI-enabled formats while keeping human voices and sources central. A preferred qualification is experience building value exchanges with creators and publishers.
Where Discover already uses AI
This is not a cold start. Google launched AI previews on trending Discover topics in the U.S., South Korea, and India in October 2025, with links back to web content. In August, Google gave users new ways to describe what they wanted more or less of in the feed. In September, Robby Stein, Google’s VP of Product for Search, announced tests of a “Dive deeper” button that opens a topic overview with related links, starting with videos.
A separate data science listing calls Discover one of the world’s largest feed products, with more than 500 million daily active users. That scale is why this hiring signal matters.
What SEO teams should monitor
Search Console already has a generative AI report for Discover. It counts impressions of links to your site in Discover’s generative AI features, broken down by page, country, and date. The Pages view lists the canonical URLs that supplied information shown in those features, which can differ from the landing page after a click.
- Check the report weekly, especially for the U.S., South Korea, and India.
- Compare AI-generated Discover impressions with standard Discover performance.
- Keep canonical URLs and authorship signals clean; Google is matching information to sources.
- Do not opt out of Search generative AI features without testing the traffic impact first.
The same Search generative AI control governs AI Overviews, AI Mode, and Discover’s generative features. Excluding a site keeps its links and content out of those surfaces, while other publishers remain available to answer the query.
The near-term play
Google has not said how much of Discover will be AI-generated. The Dive deeper tests were expected to continue in the weeks after Stein’s September post, and job listings can change or close. The right move for most teams is to treat this as a distribution shift, not a switch to flip.
A feed with 500 million daily users that can generate its own content changes the value of being source-worthy. The publishers who win will be the ones whose pages are clear enough for Google’s systems to cite and link.
Source: Search Engine Journal
