The US District Court for the District of Columbia on Friday held closing arguments in Google’s search remedies trial, a session that highlighted key questions the presiding Judge Amit Mehta might consider when making his final judgment on remedies. The trial followed Judge Mehta’s liability ruling last year finding that Google illegally maintains a monopoly over general search services and general text advertising.

The proceedings offered a window into the court’s thinking around four critical factors: the markets that remedies should apply to, how remedies might affect other markets, the proportionality of remedies, and the administrability of remedies.

Applicable Markets for Remedies

In its liability decision, the court did not define the general search market to include generative artificial intelligence products, such as ChatGPT. And the court did not find that competition in these products was harmed as a result of Google’s illegal conduct. The discussion at closing arguments suggested that imposing remedies on generative AI products would require the court to find a close nexus between them and general search.

To do this, the court is likely to consider if Google’s agreements with third-parties to distribute its search engine hinder generative AI product distribution, whether the DOJ’s proposed expansion of the definition of a search access point in its revised proposed final judgement (PFJ) to include generative AI products is justified, and whether access to Google’s unparalleled search index gives Gemini, the giant’s own generative AI product, an unsurmountable advantage.

It seems unlikely that the court can rely solely on its findings from the liability decision to answer these questions. It would have to expand on them to impose remedies on generative AI products.

The court also questioned how other proposed remedies, such as giving publishers the ability to selectively opt-out of having their web pages used in search indexing or to train Google's generative AI models (section VI B of the DOJ’s revised PFJ), and increasing advertising transparency (section VIII of the DOJ’s revised PFJ), would create more competition in general search or general search text advertising — the aim of any proposed remedies. While there seemed to be some justification for increasing advertising transparency in that it could make switching away from Google easier, there did not seem to be a direct link between the publisher opt-out remedy and competition in general search or general search text advertising during the discussion at closing arguments.

Effects on Related Markets, Like Browsers and Handsets

The exclusionary effects of Google’s distribution agreements (default payments, revenue sharing, and licensing terms) and their substantial contribution toward maintaining Google’s monopoly are clear findings in the court’s liability decision. Thus, remedies to change or prohibit these agreements and enable more competition follow naturally from those findings.

The questions the court is likely to consider are whether the DOJ’s proposal to prohibit payments to third-parties who distribute Google search (sections IV A, B and E of the DOJ’s revised PFJ) could harm competition in the browser and mobile handset markets, and in turn adversely affect competition in general search.

At closing arguments, the court asked whether, for example, giving distribution partners a year or two to wean themselves off Google’s default payments would be less disruptive, or if it would be appropriate to make exceptions for companies like Mozilla that are heavily dependent on the payments to fund their operations. At the same time, the court was clear that Google’s remedy proposals did not go far enough and that the tens of billions it pays Apple every year for default placement has influenced Apple’s decision to stay out of the general search game.

It seems clear that the court is leaning toward restricting or even prohibiting Google’s payments so that it can open up distribution for competing products, but that it would like to do so in a way that minimizes harm in other markets.

Proportionality of Remedies

At closing arguments, Judge Mehta outlined three types of remedies. First, injunctive remedies which would require Google to stop conduct that is harming competition, such as paying for default placement. Second, forward-looking remedies that stop short of structural separation, such as syndication and data sharing, which require Google to offer services it currently may not, to pry open competition. And finally, structural remedies, like forcing Google to spin off its Chrome browser.

The discussion at closing arguments suggested that the court is comfortable enough with its findings in its liability decision to impose injunctive remedies and forward-looking remedies, stopping short of structural separation with an important caveat.

The important caveat for forward-looking remedies is that these should target inputs for general search services and general text advertising that Google has acquired because of its monopoly, and the fact that potential competitors cannot replicate them without the benefits of network effects that Google enjoys.

For example, at closing arguments the court briefly went through data related to Google’s search index as specified in the US government’s proposed remedies (section VI A of the DOJ’s revised PFJ) to try and identify what was linked to Google’s monopoly and what, like the Knowledge Graph and local information, may not be related to the network effects that only the company enjoys.

The ability of potential competitors to replicate inputs independently was front and center in the discussion on the appropriate scope of syndication remedies. These remedies would allow a company to purchase organic web links and other components of a search page result from Google (section VII of the DOJ’s revised PFJ). The court and the DOJ agreed that the aim of the syndication remedy is not to enable entry by a company solely based on syndication from Google. The aim is to encourage entry by companies that will invest in their own infrastructure, like a search index. The court asked if there is a way to rate limit syndication to ensure this. This would mean the syndication licensing remedy could only be used by a licensee for a small proportion of queries, like tail queries, as suggested in an amicus brief filed by US browser company Brave.

The Administration of Remedies

Throughout closing arguments, the court raised questions about the administrability of various remedies and the need for key details of the remedies to be worked out in the final judgement and not left for later. This highlighted the court’s critical view of the role of the technical committee as proposed by the DOJ and, more broadly, the expansive role that the department and the executive branch of the US government would play in administering remedies (section X of the DOJ’s revised PFJ).

The court did not seem comfortable leaving crucial questions, such as how to balance between privacy and any potential data sharing remedies or how to decide what criteria should be used to define a qualified competitor for the technical committee to decide later. Thus, I would expect questions such as these to be answered in the court’s final judgement, which is expected in August.