AI News Google takes EU to court over data access for search engine and AI competitors

[Brownie2019]

Content source

https://cybernews.com/news/google-eu-competition-lawsuit/

It’s being asked to share its success with competitors, but Google isn’t having it.
Key takeaways:

• Google is challenging an EU order to share search data with rival search engines and AI companies.
• Google warns the demand risks user privacy, while the EU says anonymization safeguards protect users.
• If upheld, data sharing starts by January 2027, with Android changes expected by July 2027.

Google is challenging the European Commission’s demand to share its search engine data with competing search engine providers and AI companies.

In July 2026, the European Commission ordered Google to give search engine rivals and competing AI companies, such as OpenAI, access to its services, including information on rankings, entered search queries, click-through rates, and how often answers are viewed.

Competing search engine providers and AI chatbots can use this information to optimize their services to compete more effectively with Google Search.

This action was taken under the Digital Markets Act (DMA), which requires "gatekeepers" like Google to offer interoperability. The DMA aims to prevent monopolistic business practices by major tech companies and promote a more competitive, level-playing-field environment.
Opening up its services to competitors would mean that users can activate a rival AI assistant via voice commands, for example, to book a taxi or search for information.

Google has argued that sharing search data may violate European privacy laws and poses significant risks to the privacy of millions of European users.

But according to the European Commission, the measure contains robust safeguards to protect users’ privacy, including anonymization of user data, and opens up competition.

Google strongly opposes the Commission’s point of view regarding the scope of the DMA and has filed a challenge to the European Court of Justice’s (CJEU) General Court, Europe’s second-highest court.

The tech company has also applied for an interim measure at the Luxembourg-based court because Google feels that the European Commission’s order to share search engine data with rivals poses a risk of serious harm.

If upheld, Google would be required to start sharing its search data with eligible search engine providers by January 2027, with changes to its Android services expected to benefit users by July 2027.

 

[Morro]

The irony that Google is worried about our privacy.

And maybe they have a point, but if they lose, it would break their monopolistic business practices. Which would/could be a good thing I guess.

 

[Bot]

Hi Brownie2019. Google has a legitimate privacy argument here, but “share its success” makes the order sound broader than it is. The Commission is not demanding Google’s search algorithm, source code or model weights. It has issued two distinct DMA measures:

• Search data: eligible third-party search-engine providers can request anonymised ranking, query, impression and click data on fair terms. This can support search used to ground an AI chatbot, but it is not an unrestricted dump for every AI company.
• Android interoperability: rival assistants should receive access to specified Android features comparable to Gemini—for example, voice activation and actions within other apps.

The privacy issue is nevertheless serious. Removing account names and IP addresses does not automatically anonymise a search query; the query itself may contain a full name, address, medical details or even a password. Search histories have been reidentified in the past by combining distinctive queries.

However, the Commission has required considerably more than a vague promise to “anonymise” the files. Its published measures include removing direct identifiers and timestamps, breaking links between a person’s different queries, suppressing rare or unusually long queries and sensitive-looking terms, and applying group-size thresholds. Contractual controls also prohibit recipients from trying to reidentify users or link the dataset to auxiliary advertising and analytics data.

Those safeguards reduce the risk; they do not magically prove it is zero. The key legal and technical question is whether reidentification is reasonably possible with other datasets, not whether Google has removed a username column. Independent testing and strict enforcement against recipients will matter far more than either side’s press release.

My view is that opening the accumulated feedback advantage of a dominant search engine is defensible in principle, but the court should scrutinise the anonymisation design closely. Google’s commercial interest in keeping the data exclusive does not make its privacy concern false; equally, invoking privacy does not entitle it to keep every competition advantage generated at enormous scale.

The Commission’s July decision remains binding while challenged. Google has also requested an interim suspension, but unless the General Court grants it, the search-data deadline remains January 2027. That interim ruling may therefore have practical consequences well before the full appeal is decided.

Sources

• Commission provides guidance to Google for AI interoperability on ...
• Alphabet specification proceedings - Sharing of Google Search data - Digital Markets Act (DMA)