Google is recruiting machine learning engineers to generate and recommend AI-generated content in its Discover information feed, as well as senior designers to shape the user experience. The job listings include creating content tailored to users’ interests and checking its quality. They do not specify when the features will roll out or what share of the feed AI-generated content will account for.
Discover is a feed that delivers information tailored to users’ interests
Discover is a Google feed that surfaces articles, videos and other content from across the web based on users’ interests. A simple way to think about the difference is that Search lets you “enter what you want to know and look for it,” while Discover delivers information you may be interested in.
A job listing for a data science role related to Discover describes the service as having more than 500 million daily active users. This hiring effort concerns how to deliver AI-created content in a feed of that scale.
The point is not just to use AI to improve the accuracy of recommendations. The job listings also include content generation, suggesting that Google plans to integrate AI not only into how information is selected, but also into the content users read.
Toward AI that creates content, not just selects articles
The machine learning engineer listing names topic discovery, content understanding and quality management for AI-generated content among the responsibilities. The role also includes developing systems in which AI carries out multiple processes to create content for Discover and notifications.
The checks cover factual accuracy, freshness, consistency and tone. The listings also call for a system that incorporates user responses into future improvements. Senior machine learning engineering roles also involve managing the quality of AI-generated content in Discover.
The challenge is to balance content that captures users’ interest with content they can trust. Even if AI can generate engaging writing, outdated information or inaccurate explanations could undermine the value of an information service. The listings’ explicit mention of quality checks shows that this challenge is built into the development work. However, they do not reveal specific verification methods or quality standards.
The goal is an experience that supports users beyond reading information
The senior UX design role outlines a vision of making Discover feel like a concierge tailored to each person. UX is about designing not only how easy a screen is to use, but also the kind of experience people have through a service.
The job description includes helping users explore topics they care about in depth and take action based on what they learn. It also mentions working with the YouTube, Shopping, Chrome and Android teams, and expanding AI-powered formats in the feed.
These responsibilities point toward a shift from a feed where users simply read an article and leave to a service that organizes information around their goals. However, it is unclear how the services would be connected or what interactions would be possible.
The same job listing also calls for keeping human creators and sources at the center. Experience building systems that facilitate value exchange between creators and publishers is listed as a plus. As AI-powered information delivery expands, how it will coexist with the people and organizations that create the original content is also a design challenge.
AI summaries of trending topics are already available
In October 2025, Discover introduced previews in the United States, South Korea and India that briefly summarize trending topics with AI. Each preview includes links to content on the web.
A related feature has also been announced that lets users adjust their feed by describing in words what they want to see more or less of. Google has also described a test in which users can tap a “Dive deeper” button to open a short overview of a topic and links to related articles.
The hiring plans are consistent with an effort to expand these AI features. However, ideas described in job listings should not be treated as features already available. It is important to distinguish between features that have launched, those being tested and systems Google aims to develop through this hiring.
Site operators face effects that impressions alone cannot show
Search Console for site owners includes a report showing how many times links to their sites appeared in Discover’s generative AI features. Data can be grouped by page, country and date. Page-level impressions show the canonical URL used for information in an AI feature, which may differ from the URL users visit after clicking.
However, the only metric described in the documentation is impressions. There is no breakdown by individual AI feature, and Search Labs experiments are excluded. Sites with few impressions may not see the report at all.
For site operators, having a site used in an AI explanation and actually attracting readers are separate issues. It is difficult to determine the impact on traffic or business from this report alone.
Placement in Discover’s generative AI features is governed by the generative AI controls for Search, which also apply to AI Overviews and AI Mode. Sites that opt out will no longer have their links and content used in Search’s generative AI features. The job listing does not specify a launch schedule for AI-generated content or what share of Discover’s content it will make up.
Comments