The European Union's antitrust regulator formally announced on July 17, 2026, that it would impose new compliance obligations on Google under the Digital Markets Act (DMA). According to the latest ruling, Google must share its core search ranking data with third-party search engines and app developers, and open up AI function interfaces within the Android system, including modules such as Gemini Assistant and smart search suggestions.
This decision stems from the EU's long-standing concerns over Google's dominant position in the digital market. Since the DMA took effect in 2023, Google has been designated as a "gatekeeper" company, required to adhere to a series of fair competition rules. However, the intensity of these new regulations far exceeds previous rectification requirements—not only mandating data portability but also directly intervening in the distribution methods of AI services.
Data Sharing: Opening the Underlying Logic of Search Rankings
Under the EU's requirements, Google must provide competitors with algorithm parameters used for web search, shopping search, and local search rankings, to the extent technically feasible. This means small and medium-sized search engines like DuckDuckGo and Ecosia will be able to obtain data support at the same level as Google Search, thereby optimizing the quality of their own results. The EU states that this move aims to break down "data barriers" and promote diversified competition in the search market.
Google responded in an official blog post, stating that search ranking data involves complex user behavior modeling and anti-spam systems, and that direct sharing could expose core trade secrets and increase the risk of malicious manipulation of search results. "We believe regulators have underestimated the potential threats these data pose to privacy and security, especially when they fall into the hands of small companies lacking security architectures," a Google spokesperson said.
"Forcing the sharing of search data is like asking Coca-Cola to disclose its recipe—it may increase competition in the short term, but it will undermine innovation incentives in the long run."
Opening AI: A New Shift in the Android Ecosystem
More attention has been drawn to the EU's requirement to open Android AI functions. Google must allow third-party apps to directly call the AI model interfaces of the Android system, including real-time speech recognition, image generation, and smart reply generation. This means device manufacturers like Samsung and Xiaomi, as well as independent developers, will be able to bypass Google's service framework and deeply integrate AI capabilities without being tied to the Google Play ecosystem.
Analysts point out that this will have a profound impact on the Android AI ecosystem. Currently, Google controls the entry point to Android AI through the Gemini API, but opening it up could lead to alternative solutions akin to "de-Googling." However, this also introduces security risks: open interfaces could be exploited by malware or lead to reverse engineering of AI models.
Editor's Note: The Eternal Tug-of-War Between Regulation and Innovation
The EU's move once again highlights the consistent logic of digital regulation: weakening platform monopolies through mandatory interoperability. However, similar cases—such as Microsoft being forced to open the Windows API in 2009—ultimately did not significantly alter the market landscape, instead causing competitors to fall into compatibility traps. For the AI field, openness may accelerate technology diffusion, but it could also erode user trust due to frequent security vulnerabilities.
Google's privacy warnings are not unfounded. Search data and AI models are core assets accumulated by Google over a decade, and mandatory sharing does carry the risk of misuse. On the other hand, without regulation, gatekeeper companies could indefinitely prolong data closure. Finding a balance between "open innovation" and "secure control" will be a central issue in global digital policy for years to come.
As of press time, Google has not announced whether it will file an appeal. But it is foreseeable that this litigation will be protracted and could become a landmark case in the history of AI regulation.
This article is compiled from Ars Technica
© 2026 Winzheng.com 赢政天下 | 转载请注明来源并附原文链接