Google's AI search increasingly sends users to YouTube and its own services instead of third-party sites

Google's generative AI search is turning into a self-promotion machine. A new investigation found that AI Overview and the company's other AI tools systematically direct users to Alphabet's own services (YouTube, Google Search, Google Maps, Google Shopping) at the expense of independent publications and third-party resources.

Google launched AI Overview back in 2023 under the name Search Generative Experience, and in 2024 the feature reached hundreds of millions of users worldwide. The idea is to save time: instead of clicking through several links, people get a ready-made summary right on the search page.

But an analysis of what this AI actually links to revealed a consistent pattern. YouTube, Google search pages and other Alphabet services regularly show up first in the list of cited sources. Independent news sites, specialist blogs and niche expert resources are mentioned far less often, and as a result get fewer clicks.

For the media industry this is not an abstract technical question. In 2023–2024, many publications already reported drops of 20–50% in referral traffic from Google search. AI Overview has accelerated the trend: users get the answer on the results page and don't click through to the source site. It now turns out that even when the AI does include a link, it favors platforms inside the company's own ecosystem.

YouTube is a key element of the strategy. Alphabet's video platform is among the world's three biggest sites and has long competed with traditional media for audience attention. When AI search recommends watching a video on YouTube instead of reading an article on an independent outlet, it hits the monetization of the media business directly.

Google denies deliberately promoting its own services. Company representatives attribute the frequency of YouTube citations to the quality and authority of its content, noting that the platform really is one of the most-cited resources on the web. Critics reject this argument. In their view, the algorithm itself is trained on data in which Google services are disproportionately represented, which means the bias is built into the model's architecture.

The regulatory context sharpens the issue. In the US, the antitrust case against Google continues: the Department of Justice accuses the company of abusing its dominant position in the search market. European regulators have already fined Google billions of euros for similar behavior, promoting its own services in search results to the detriment of competitors. Now the same logic is being reproduced in AI search, and regulators will have to decide whether the old rules apply to the new format.

What is happening exposes a fundamental contradiction in Google's position. The company claims the role of a neutral information aggregator while remaining the internet's largest producer of content and services. When these roles are combined in a single AI product, its neutrality inevitably comes into question.

While regulators search for an answer, independent publishers keep losing traffic, bit by bit but systematically.