After winning a fight to get Google’s search business declared an unlawful monopoly, the Department of Justice has released its initial proposal for how it’s thinking about limiting Google’s dominance — including breaking up the company.

The government is asking Judge Amit Mehta for four different types of remedies to Google’s anticompetitive power in search engines. They include behavioral remedies, or changes to business practices, as well as structural remedies, which would break up Google. And they’re focused particularly on futureproofing the search industry for the rise of generative AI. While AI might not be a substitute for search engines, the DOJ warns, it “will likely become an important feature of the evolving search industry.” And it aims to prevent Google from using its power in the industry to regain unfair control.

The government sees four areas where it can constrain Google’s power. In these, it’s asking Judge Mehta to limit the kinds of contracts Google can negotiate, require rules for nondiscrimination and interoperability, and change the structure of its business. “Fully remedying these harms requires not only ending Google’s control of distribution today, but also ensuring Google cannot control the distribution of tomorrow,” the government says. Google, for its part, calls the government’s proposals “radical” and believes they’re “signaling requests that go far beyond the specific legal issues in this case.”

What does that mean for Google’s competitors and the average user? Let’s take a look.

Search distribution and revenue sharing

Google’s search engine comes preloaded on many phones and is the default on most major web browsers, often because of revenue sharing deals that pay for that placement. The DOJ argued in court that consumers are highly unlikely to switch to a competitor, and Google’s commercial partners also have little incentive to do so while they’re getting paid. So, the DOJ says “undoing” Google’s impact on distribution is the “starting point for addressing Google’s unlawful conduct.”

The government says it’s considering restrictions on the contracts Google can strike with phone makers and browser companies, particularly agreements to make Google Search a default or have it preinstalled.

Google doesn’t just strike deals with other companies, though — it promotes its search engine and AI business with an array of other Google products, including Chrome, Android, and the Play Store. The DOJ argues this limits the available channels and incentives for rivals to compete. It didn’t offer too many details about how to fix this but said it’s considering both behavioral and structural remedies — in other words, a breakup.

And finally, a group of states that filed a related lawsuit want to address a final problem: users not realizing they’ve got other options. They’re considering ways Google could support education campaigns that inform people about competing search engines.

Accumulation and use of data

At trial, the government argued that Google creates a self-reinforcing cycle of dominance through users’ query data. Essentially, the more queries a search engine gets, the more information it has about what constitutes a useful answer, and the better the search engine becomes. Because Google’s rivals don’t have the same access to distribution channels that Google does, the DOJ argued, Google is able to siphon away most of these queries, making it much more difficult for rivals to improve their products and effectively compete.

The government says it wants to try to “offset” that advantage, potentially by forcing Google to make certain information and aspects of its product available to rivals. That could include things like data, indexes, and models used in “AI-assisted search features” and ranking signals Google uses in search.

The DOJ says it’s “mindful of potential user privacy concerns” that making Google share data would raise. “However, genuine privacy concerns must be distinguished from pretextual arguments to maintain market position or deny scale to rivals.” Privacy tradeoffs are a common defense from big tech firms against opening up their ecosystems, but the government is making clear that it doesn’t think they’re a be-all and end-all. It’s considering prohibiting Google from using data that “cannot be effectively shared with others on the basis of privacy concerns.”

Generation and display of search results

The government is also concerned with “new and developing features of general search,” particularly generative AI. In this field, it argues much of Google’s power stems from scraping data from sites with “little-to-no bargaining power against Google’s monopoly.” These sites may not want their work used to train AI models, but they “cannot risk retaliation or exclusion from Google” by blocking its crawlers altogether. As a result, the DOJ is considering requiring that Google let sites opt in to search engine inclusion while opting out of inclusion in its AI tools.

Advertising scale and monetization

In addition to the general search market, Mehta found that Google has monopoly power in the market for general search text ads: the sponsored text links that appear when you enter a query. (This is different from the market at issue in Google’s second legal battle with the DOJ, which dealt with the tools used to serve ads on publishers’ sites across the internet.) Mehta found that Google does not consider rivals’ prices in figuring out its own for this product — something he says only a monopolist with no meaningful competition could do.

To fix this, the government says it’s considering remedies that would “create more competition and lower the barriers to entry, which currently require rivals to enter multiple markets at scale.” That could include addressing Google’s use of AI to protect its monopoly power in this market, it says. It’s also looking at licensing or syndicating Google’s ad feed separately from its search results and requiring certain kinds of information Google should provide to advertisers about their performance.