What Will DoJ Require of Google?

Google is now in the “remedies” phase of an antitrust lawsuit filed by the U.S. Department of Justice, as Google has been found to be a monopolist in the search market. 

So what matters now are the remedies. DoJ has been floating remedies including

• Prohibiting Google's exclusive default search agreements, such as the multibillion-dollar deal with Apple to be the default search engine on Safari

• Forcing Google to divest its Chrome browser

• Requiring Google to share certain search data with competitors to level the playing field. ​

A final ruling on the remedies is expected by August 2025. ​An important issue is the relevance of past remedies imposed on Microsoft in the past. Basically, the case challenged Microsoft's bundling of its Internet Explorer web browser with the Windows operating system. The case culminated in a 2002 settlement that focused on behavioral remedies.

Key behavioral remedies included:

• API disclosure to third-party developers

• OEM flexibility, allowing original equipment manufacturers to install non-Microsoft middleware (Netscape Navigator)

• Set non-Microsoft middleware as defaults (a different default browser)

• Remove or hide access to Microsoft's middleware

Also, Microsoft was required to license Windows on uniform terms, meaning it could not provide preferential treatment to partners that agreed to exclude competitors.

If Google tries to follow Microsoft's playbook by arguing for behavioral remedies only, it is likely to  face a more skeptical reception than Microsoft did in the early 2000s. Regulators today are more likely to argue that behavioral remedies alone will not work.

In particular, the divesting of the Chrome browser has been proposed. That could have implications for artificial intelligence apps for several reasons, including the role played by the Chrome browser in driving traffic to Google’s important search engine. 

Losing the Chrome browser would likely require Google to do more cross-platform deals. 

For AI assistants, search provides real-time information beyond training data cutoffs, allowing answers including current information. Search also allows models to retrieve specific information on demand. Integrated search also helps models overcome hallucination issues by verifying facts. 

Search might also be a platform for creating new “super apps” that offer many functions within a single application. 

So the actual remedies will matter.