Google's "Preferred Sources" feature is a free pass for more garbage in search
Key Points
Google has introduced a "Preferred Sources" feature that lets users manually mark journalistic outlets they want to see more often in search results.
But the company has spent decades algorithmically tracking source quality and user preferences, making a manual tool for this purpose entirely unnecessary if the actual goal were better search results. Instead, the feature shifts responsibility to users and serves as a ready-made defense against European regulators.
And as AI-generated answers increasingly replace links to external sites, Google wants to control which sources feed its systems, favoring compliant partners over independent publishers who might demand compensation or assert their rights.
Google lets users manually choose which sources show up more often in search, supposedly to support quality journalism. The holes in that argument are obvious.
Google, of all companies. The company that has spent decades saying it understands what users want better than anyone else. The company with perhaps the most comprehensive trove of user data in human history.
And now, in the age of AI, that same company supposedly needs users to manually mark news outlets as "preferred sources" so reputable journalism shows up more often in search results.
That doesn't hold up. If Google wanted to prioritize high-quality sources, it already could. The company knows which outlets are reliable, which sources users click on, and which content is produced by editors and reporters. So why is Google suddenly handing that decision off to users?
This is about controlling the source layer
The real answer is simpler—and, of course, uglier: Google wants more control over which sources feed its AI search and which ones stay visible. That's especially useful when fewer of those sources push back over copyright, compensation, or reach.
Google's "Preferred Sources" feature is a free pass for more garbage in search
Key Points
Google has introduced a "Preferred Sources" feature that lets users manually mark journalistic outlets they want to see more often in search results.
But the company has spent decades algorithmically tracking source quality and user preferences, making a manual tool for this purpose entirely unnecessary if the actual goal were better search results. Instead, the feature shifts responsibility to users and serves as a ready-made defense against European regulators.
And as AI-generated answers increasingly replace links to external sites, Google wants to control which sources feed its systems, favoring compliant partners over independent publishers who might demand compensation or assert their rights.
Traditional Google search has long been a trade-off. Google monetized the search query, but it still sent users to external websites. With AI-generated answers, that relationship is changing. Google still uses the content, but it increasingly keeps users inside its own interface. Sources become raw material.
That creates a new incentive. High-quality original sources are often the best results for users, but they're inconvenient for Google: they're editorial, monetized, and able to push back with legal claims. Automated or semi-automated AI spam pages are easier to handle. They're happy with scraps of traffic, and they don't complain. Whether their visibility is intentional or collateral damage matters less than the result. The effect works in Google's favor, because hardly anyone clicks source links in AI responses anyway.
Google is also locking down content where the terms are easier to manage. For example, the company struck a deal with Reddit to use Reddit content for AI training. Around the same time, Reddit surged in Google Search results. That's far more predictable for Google than dealing with an open network of publishers, rights holders, and potential lawsuits.
"Preferred Sources" fits neatly into this logic. The feature shifts responsibility instead of fixing the problem of bad search results. Google knows that only a tiny fraction of users will ever manually add preferred sources. But the mere existence of the option is useful: users could choose different sources.
That effectively turns not using the feature into consent to Google's defaults. If you don't set preferred sources, you get whatever Google serves up. And if that means AI spam, convenient partners, or some other substitute source, Google can say the user could have changed the selection. It's a clean way to sideline both users and reputable publishers.
Google gets a convenient answer for regulators
The strategy is also political. Google is under pressure on several fronts in Europe. The Digital Services Act (DSA) requires major search engines to be transparent about their recommendation algorithms and to reduce risks to media pluralism. Italy's media regulator AGCOM has already asked the European Commission to examine Google's AI Overviews under the DSA.
The Digital Markets Act (DMA) covers Google Search as a gatekeeper service. Among other things, it bans self-preferencing in rankings and requires non-discriminatory conditions. Separately, the European Commission has been investigating since December 2025 whether Google is using publishers' content for AI purposes without appropriate compensation under Article 102 TFEU.
Against that backdrop, "Preferred Sources" looks like a ready-made argument for regulators and lawmakers. Going forward, Google can say users can influence source selection themselves. Whether people actually use that option at any meaningful scale is beside the point. What matters is that Google can point to it. In practice, source selection almost always stays with Google.
The open web is no longer Google's growth business
If you want to know how much Google still values the open web, look at the numbers. Google's ad business is growing quickly, but not where the open web lives. Search and YouTube are growing. The Google Network, where Google places ads on external websites, is shrinking since 2022.
Key figure Q1 2025 Q1 2026 Google Search & other USD 50.702 billion USD 60.399 billion YouTube ads USD 8.927 billion USD 9.883 billion Google Network USD 7.256 billion USD 6.971 billion Google advertising total USD 66.885 billion USD 77.253 billion
The trend is clear: Google is making more money from its own products and less from ads on third-party websites. Google's lawyers were accidentally candid during an ad rights dispute in fall 2025, referring to the "rapid decline" of the open web. They meant Google's shrinking display ad revenue. But that's exactly the point.
If the open web matters less to Google's business, Google has less reason to send users there. Instead, the company is pulling them deeper into its own ecosystem through AI features. The worse, more confusing, and more interchangeable the web looks, the easier it is to keep users inside Google's own products.
"People love our AI experiences like AI Mode and AI Overviews, and they're coming back to search more," Google CEO Sundar Pichai said when discussing the latest financial results. From Google's perspective, that makes sense: when answers appear directly in search, attention stays with Google. The source becomes a supplier, not a destination.
This playbook isn't new. Facebook, Instagram, LinkedIn, and X have been devaluing external links for years to keep users on their own platforms. Now Google is moving in the same direction. External websites are only truly useful to Google if they feed its products or can be monetized through ad clicks. Even click-based ads can benefit when organic search gets worse: users who can't find anything useful in the organic results are more likely to click paid placements, especially if the line between an ad and a search result is hard to spot.
Publishers should stop giving Google cover
Many publishers have already started using big buttons to ask readers to add their sites to Google's "Preferred Sources." That's understandable in the short term. If you're losing reach, you grab every lever you can.
Strategically, it's a mistake. Because this is exactly how Google's cover works. The company builds a feature almost nobody will use, and publishers do the work of driving users to it. In the process, they help Google turn a theoretical control option into something that looks like a serious product.
Any publisher suing Google at the same time should ask whether it makes sense to immediately build every new Google feature into its own user flow. That only legitimizes the mechanism publishers should be challenging: Google decides which sources are visible, then points to a manual user option as soon as that decision comes under scrutiny.
Google has given itself cover with "Preferred Sources." The industry shouldn't help prop it up.
When I asked Google why it needs a manual feature like "Preferred Sources" after decades of algorithmically tracking source quality and user preferences, it didn't answer the question.
Google lets users manually choose which sources show up more often in search, supposedly to support quality journalism. The holes in that argument are obvious.
Google, of all companies. The company that has spent decades saying it understands what users want better than anyone else. The company with perhaps the most comprehensive trove of user data in human history.
And now, in the age of AI, that same company supposedly needs users to manually mark news outlets as "preferred sources" so reputable journalism shows up more often in search results.
That doesn't hold up. If Google wanted to prioritize high-quality sources, it already could. The company knows which outlets are reliable, which sources users click on, and which content is produced by editors and reporters. So why is Google suddenly handing that decision off to users?
This is about controlling the source layer
The real answer is simpler—and, of course, uglier: Google wants more control over which sources feed its AI search and which ones stay visible. That's especially useful when fewer of those sources push back over copyright, compensation, or reach.
Traditional Google search has long been a trade-off. Google monetized the search query, but it still sent users to external websites. With AI-generated answers, that relationship is changing. Google still uses the content, but it increasingly keeps users inside its own interface. Sources become raw material.
That creates a new incentive. High-quality original sources are often the best results for users, but they're inconvenient for Google: they're editorial, monetized, and able to push back with legal claims. Automated or semi-automated AI spam pages are easier to handle. They're happy with scraps of traffic, and they don't complain. Whether their visibility is intentional or collateral damage matters less than the result. The effect works in Google's favor, because hardly anyone clicks source links in AI responses anyway.
Google is also locking down content where the terms are easier to manage. For example, the company struck a deal with Reddit to use Reddit content for AI training. Around the same time, Reddit surged in Google Search results. That's far more predictable for Google than dealing with an open network of publishers, rights holders, and potential lawsuits.
"Preferred Sources" fits neatly into this logic. The feature shifts responsibility instead of fixing the problem of bad search results. Google knows that only a tiny fraction of users will ever manually add preferred sources. But the mere existence of the option is useful: users could choose different sources.
That effectively turns not using the feature into consent to Google's defaults. If you don't set preferred sources, you get whatever Google serves up. And if that means AI spam, convenient partners, or some other substitute source, Google can say the user could have changed the selection. It's a clean way to sideline both users and reputable publishers.
Google gets a convenient answer for regulators
The strategy is also political. Google is under pressure on several fronts in Europe. The Digital Services Act (DSA) requires major search engines to be transparent about their recommendation algorithms and to reduce risks to media pluralism. Italy's media regulator AGCOM has already asked the European Commission to examine Google's AI Overviews under the DSA.
The Digital Markets Act (DMA) covers Google Search as a gatekeeper service. Among other things, it bans self-preferencing in rankings and requires non-discriminatory conditions. Separately, the European Commission has been investigating since December 2025 whether Google is using publishers' content for AI purposes without appropriate compensation under Article 102 TFEU.
Against that backdrop, "Preferred Sources" looks like a ready-made argument for regulators and lawmakers. Going forward, Google can say users can influence source selection themselves. Whether people actually use that option at any meaningful scale is beside the point. What matters is that Google can point to it. In practice, source selection almost always stays with Google.
The open web is no longer Google's growth business
If you want to know how much Google still values the open web, look at the numbers. Google's ad business is growing quickly, but not where the open web lives. Search and YouTube are growing. The Google Network, where Google places ads on external websites, is shrinking since 2022.
Key figure Q1 2025 Q1 2026 Google Search & other USD 50.702 billion USD 60.399 billion YouTube ads USD 8.927 billion USD 9.883 billion Google Network USD 7.256 billion USD 6.971 billion Google advertising total USD 66.885 billion USD 77.253 billion
The trend is clear: Google is making more money from its own products and less from ads on third-party websites. Google's lawyers were accidentally candid during an ad rights dispute in fall 2025, referring to the "rapid decline" of the open web. They meant Google's shrinking display ad revenue. But that's exactly the point.
If the open web matters less to Google's business, Google has less reason to send users there. Instead, the company is pulling them deeper into its own ecosystem through AI features. The worse, more confusing, and more interchangeable the web looks, the easier it is to keep users inside Google's own products.
"People love our AI experiences like AI Mode and AI Overviews, and they're coming back to search more," Google CEO Sundar Pichai said when discussing the latest financial results. From Google's perspective, that makes sense: when answers appear directly in search, attention stays with Google. The source becomes a supplier, not a destination.
This playbook isn't new. Facebook, Instagram, LinkedIn, and X have been devaluing external links for years to keep users on their own platforms. Now Google is moving in the same direction. External websites are only truly useful to Google if they feed its products or can be monetized through ad clicks. Even click-based ads can benefit when organic search gets worse: users who can't find anything useful in the organic results are more likely to click paid placements, especially if the line between an ad and a search result is hard to spot.
Publishers should stop giving Google cover
Many publishers have already started using big buttons to ask readers to add their sites to Google's "Preferred Sources." That's understandable in the short term. If you're losing reach, you grab every lever you can.
Strategically, it's a mistake. Because this is exactly how Google's cover works. The company builds a feature almost nobody will use, and publishers do the work of driving users to it. In the process, they help Google turn a theoretical control option into something that looks like a serious product.
Any publisher suing Google at the same time should ask whether it makes sense to immediately build every new Google feature into its own user flow. That only legitimizes the mechanism publishers should be challenging: Google decides which sources are visible, then points to a manual user option as soon as that decision comes under scrutiny.
Google has given itself cover with "Preferred Sources." The industry shouldn't help prop it up.
When I asked Google why it needs a manual feature like "Preferred Sources" after decades of algorithmically tracking source quality and user preferences, it didn't answer the question.