Bryan Koenig
December 26, 2025
OpenAI Opposes 'Cookie-Cutter' Google Search Fixes
6 min
AI-made summary
- OpenAI submitted an amicus brief in the Justice Department's antitrust case against Google, urging the D.C
- federal court to adopt flexible remedies for Google's search monopoly
- OpenAI opposed Google's proposal to limit syndicated search results to 'ten blue links,' arguing it would hinder innovation and generative AI products like ChatGPT
- DuckDuckGo also filed a brief supporting broader data sharing and privacy protections
- Judge Mehta is considering how to structure remedies to foster competition among search providers.
OpenAI waded into the Justice Department's case against Google's search monopoly Friday to urge the D.C. federal judge to apply flexibility to mandates requiring Google to syndicate its search results to would-be rivals, arguing that permitting Google's more rigid "ten blue links" proposal would stifle "innovative uses."
With U.S. District Judge Amit P. Mehta weighing exactly how to craft mandates requiring Google to prop up rivals with data and syndicated search results, OpenAI OpCo LLC said in a proposed amicus brief that the court should side with the U.S. Department of Justice's approach, instead of Google restrictions that "would effectively gut these remedies" by limiting competitors' display of syndicated search results to 10 blue links.
Google's proposed restrictions "would force qualified competitors to adopt a single mode of displaying search results ... and thus bar innovative uses of the results," said OpenAI, best known for its ChatGPT product. "Google's proposed cookie-cutter approach to displaying search results is simply not how GenAI products (like ChatGPT) use search syndication, and it is telling that Google itself does not use the ten blue links for Google's own AI search product named 'AI Overviews.'"
OpenAI advocated instead for the DOJ's call for "the least restrictive terms" Google currently offers in existing search syndication deals, in order to use what it obtains in a way that meshes with ChatGPT's conversational, narrative responses to user queries.
The goal of the remedies is to open up competition that Judge Mehta found last year had been closed off by Google's effectively exclusive distribution contracts with Mozilla, Apple and Samsung, under which it paid billions of dollars annually in shared search advertising revenue to be made the default search engine on their browsers and devices.
While Judge Mehta in September rejected calls by the DOJ and its state partners to divest the Chrome browser or ban those default contracts and payments entirely, he decided Google should prop up current and would-be search rivals left behind by the default agreement lockout and the massive data advantage the lockout ultimately conferred. His Sept. 2 order explicitly envisioned treating generative artificial intelligence companies like OpenAI as qualified competitors eligible to receive Google's data and syndicated results.
That ruling teed up further briefing on how to craft the final judgment that Judge Mehta hopes to issue relatively quickly, where he'll decide between Google's push for the most restrictive remedies it can get and calls by the DOJ to give qualified competitors as much of a boost as possible.
OpenAI is one of two likely data recipients to weigh in so far, alongside privacy-focused search rival DuckDuckGo, which submitted its own proposed amicus brief Oct. 1, before Judge Mehta held an early October hearing on the competing proposals.
In Friday's brief, OpenAI opposed Google's calls for permitting syndication to present the search results only to the searching user, noting that even Google's own AI Overviews doesn't answer queries with the 10 blue links. Instead, OpenAI noted that as it pushes its way directly into search — instead of ChatGPT responses frozen in time to the most recent data used to train the model — it's not interested in providing the same kind of search experience that users now get with Google.
OpenAI further called for syndication provided on terms no less favorable than the best terms Google currently offers, warning that Google's proposal for "ordinary" commercial terms would be based on "a handful of outdated agreements negotiated and executed by Google with the benefit of its monopoly power."
Nor should Google be able to impose "artificial and ambiguous limitations" on how its user-side data can be shared, according to the brief. Noting that it lacks "sufficient scale of click-and-query data on its own to build a search technology that can effectively compete against Google," OpenAI said the fix should allow use to develop its own products.
OpenAI said that both it and the DOJ read Google's data sharing proposal as permitting only query response and barring qualified competitors from using the data to build their own index, "an artificial limitation that would neuter the remedy, permitting it to maintain its long-held monopoly in search."
DuckDuckGo's own amicus brief said that an envisioned technical committee needs "broad authority" to ensure Google complies with the mandates, that syndication licenses should be available on five-year terms that can extend beyond the remedies' six-year term — to prevent Google from slow-walking negotiations to limit the length of sharing — and that Google shouldn't be able to require recipients of end-user data to provide extensive information back to the search giant.
On that last point, DuckDuckGo called, like OpenAI, for user data sharing on terms at least as favorable as Google's current search syndication. Noting its own focus on privacy, with a search engine that collects no user data, DuckDuckGo warned against Google's calls for rivals to share data with it under "ordinary commercial terms."
DuckDuckGo said it knows of at least one private search engine that syndicates Google search results on terms that protect users' privacy.
"To the extent that Google deems this provision (and potentially similar provisions in other syndication contracts) non-'ordinary,' it should not be permitted to withhold similar provisions from qualified competitors," DuckDuckGo said. "To do so would force DuckDuckGo and other private search engines to choose between forgoing the syndication remedy … and forgoing their privacy focus, the key differentiation from Google's product that the court has acknowledged will be necessary for them to compete."
Representatives for OpenAI and Google could not immediately respond Friday to requests for comment. Law360 reached out to the DOJ but it is subject to the government shutdown.
Future disputes in the case will not be limited to the ultimate injunction. Google has vowed to appeal, as it has in the separate DOJ case over its advertising placement technology monopolies, where the parties just wrapped an 11-day remedies phase trial on the government effort for a breakup of the business. It's also possible the DOJ will file a cross-appeal of the parts of Judge Mehta's remedies decision it doesn't like.
DOJ officials have so far said there's been no decision on whether to appeal.
OpenAI is represented by Ashok Ramani and Serge A. Voro of Davis Polk & Wardwell LLP.
DuckDuckGo is represented by Melissa H. Maxman and Ronald F. Wick of Cohen & Gresser LLP.
The federal government is represented by David Dahlquist, Adam Severt, Veronica Onyema, Travis Chapman, Diana Aguilar, Sarah Bartels, Grant Fergusson, Kerrie Freeborn, Meagan Glynn, Richard Cameron Gower, Karl Herrmann, Ian Hoffman, John Hogan, Elizabeth Jensen, Ryan Karr, Claire Maddox, Michael McLellan, Keane Nowlan, Andrew Tisinger, Jennifer Wamsley and Catharine Wright of the U.S. Department of Justice's Antitrust Division.
The states are represented by their respective attorneys general and William F. Cavanaugh Jr. of Patterson Belknap Webb & Tyler LLP.
Google is represented by John E. Schmidtlein, Benjamin M. Greenblum, Colette T. Connor, Kenneth C. Smurzynski, Graham W. Safty, Christopher Yeager, Gloria K. Maier and Aaron P. Maurer of Williams & Connolly LLP, Michael S. Sommer and Franklin M. Rubinstein of Wilson Sonsini Goodrich & Rosati PC and Mark S. Popofsky and Matthew L. McGinnis of Ropes & Gray LLP.
The cases are U.S. et al. v. Google LLC, case number 1:20-cv-03010, and Colorado et al. v. Google LLC, case number 1:20-cv-03715, in the U.S. District Court for the District of Columbia.
Article Author
Bryan Koenig
The Sponsor
