Written by Julien Ricciarelli-Bonnal
29 September 2026
The Essentials
Google is challenging two European decisions that would force it to open more of its ecosystem to competitors, including artificial intelligence companies. One of the most sensitive issues concerns data generated by Google Search. Despite the language sometimes used around the dispute, the EU is not demanding access to individual users’ personal search histories. It wants Google to share anonymised data about queries, clicks, views and rankings. Behind this technical battle lies a much larger issue: the traces produced by billions of searches have become a strategic resource in the competition between Google, OpenAI and the next generation of AI-powered search engines.
For more than twenty years, a Google search had an obvious value for the person making it: finding information, a company, a product or an answer. For Google, however, it also had another value. Every query, every click and every result opened generated signals that could gradually improve the search engine itself.
Taken individually, those actions look insignificant. At the scale of billions of searches, they become an extraordinary raw material. They reveal which formulations lead to useful results, which links genuinely satisfy an intention, which pages are ignored and how behaviour changes when the search engine modifies its rankings.
This accumulated advantage is precisely what the European Union now wants to make partially accessible to Google’s competitors. Google is challenging the measures before the General Court of the European Union, arguing that the requirements imposed under the Digital Markets Act could threaten user privacy.
Behind what appears to be a highly technical regulatory dispute lies a much simpler question: who should benefit from the advantage created by decades of searches performed by internet users?
Europe is not really asking for your personal Google history
The language surrounding this battle deserves clarification. Google argues that the European decisions will force it to share Europeans’ private searches and says the proposed anonymisation measures may not provide sufficient protection. The image is powerful: medical searches, financial concerns, relationship problems or professional questions potentially ending up in the hands of rival technology companies.
The European Commission describes something significantly different. Its decision requires Google to provide certain competing search services with access to data generated through Google Search, including information about queries, clicks, views and rankings. AI assistants that provide search functionality may also qualify.
But the Commission explicitly excludes account information and individual user histories. The data must undergo several transformations before being shared, including the removal of certain rare or unusually long queries, reduced geographical precision, the absence of exact timestamps and the aggregation of some interaction information.
In other words, OpenAI or another eligible competitor would not receive a file showing that one identifiable person searched for a medical symptom on Monday night and then looked for a lawyer the following morning. The aim is to give competitors access to part of the vast pool of signals Google uses to understand how people search and interact with results.
That distinction matters, but it does not eliminate the privacy debate. Google argues that even transformed data can carry risks of re-identification or reveal sensitive information. The Commission believes the technical protections, contractual restrictions and auditing mechanisms are sufficient. The disagreement therefore concerns not whether sensitive search behaviour exists, but whether it can be made useful to competitors without exposing the people who generated it.
A Google search has become raw material for AI
Why is Europe so interested in this data? Because Google’s position does not rest only on a powerful algorithm or enormous infrastructure. It also rests on decades of interaction with users.
Every query provides a small piece of information about human language. Every click offers a signal about relevance. Every abandoned result may indicate that the answer was unsatisfactory. When those signals accumulate at a scale no new search engine can easily reproduce, they become extremely valuable for improving query understanding, ranking, suggestions and information retrieval.
AI assistants are now moving directly into this territory. ChatGPT, Gemini, Perplexity and their competitors are no longer limited to producing text from information learned during training. They increasingly search the live web, identify sources and construct answers through retrieval systems.
User behaviour is changing at the same time. As Google searches become longer and more conversational under the influence of AI-powered answers, the ability to understand complex intentions rather than simply match keywords is becoming even more valuable.
The Commission’s reasoning is therefore straightforward: if new search competitors are permanently denied access to the kinds of behavioural signals that helped Google improve its own engine, the barrier to entry may become almost impossible to overcome.
The AI race is consequently being fought over assets far less spectacular than models themselves. Having the strongest model matters. Having data that reveals what people search for, what they select and what they ultimately consider relevant may matter just as much.
OpenAI could benefit from Google’s data, but not for everything
The possibility that companies such as OpenAI could benefit makes the dispute considerably more interesting. The European Commission says that AI assistants offering search functionality can qualify for access to the mechanism. They could therefore use some Google Search data to improve their own search systems.
That does not mean Google would be required to provide data for the direct training of OpenAI’s next general-purpose model. The European decision limits how the information can be used. It is intended to support search-related functions such as query understanding, ranking, retrieval and indexing.
The Commission explicitly prohibits broader uses including the training of general-purpose AI models, advertising profiling and the improvement of unrelated services. Europe is therefore attempting something unusually precise: opening part of Google’s informational advantage without turning user-generated search data into freely exploitable material for any technology company.
Maintaining that boundary may become increasingly difficult. Search engines and AI assistants are progressively merging into the same interface, making it harder to separate technology used to improve search from technology used to improve the assistant as a whole. A system that understands a request, selects sources and generates a synthesis relies on several deeply interconnected components.
The conflict is therefore already larger than Google. It raises a much broader question about how Europe intends to organise the circulation of data in an economy where search engines, personal assistants and artificial intelligence are gradually becoming the same product.
Our data is becoming an instrument of competition policy
There is a striking paradox here. For years, most debates about personal data focused on restricting how large platforms could collect and exploit it. The Digital Markets Act introduces a different logic: some data accumulated by a dominant platform may also become a tool for encouraging competition.
That creates a direct tension between two objectives. The more strongly data is protected and anonymised, the lower the risk to users, but potentially the less useful it becomes to competitors. The more precise it remains, the greater its technological value, but the more sensitive the privacy implications become.
Google clearly has a major commercial interest in preserving exclusive access to the value of its search data, which does not automatically invalidate its privacy concerns. The European Commission, meanwhile, has a strong interest in demonstrating that its safeguards are sufficient to make the opening both legally and politically defensible.
For companies, this battle is also a reminder that AI governance, data use and organisational priorities increasingly need to be considered together. Data created through everyday activity can acquire an entirely new strategic value once systems become capable of learning from behaviour, reconstructing intentions and improving their decisions from accumulated patterns.
The artificial intelligence race is usually described through billions invested in models, data centres and chips. The dispute between Google and Europe shows that part of the competition is also being fought over something users produce almost without noticing: the traces left behind when they search for an answer.
For twenty years, those traces primarily strengthened Google. Europe now wants part of their value to help competitors emerge. Google argues that opening them could go too far.
For once, the billions of searches performed every day are no longer merely fuel for a search engine. They are becoming an instrument of power in the battle to build the one that comes next.
If artificial intelligence is changing the value of your data, tools or processes, we can help you identify the uses worth prioritising and build the rules needed to govern them.
Written by Julien Ricciarelli-Bonnal
29 September 2026
