虎嗅

**Google Forced to Open Doors: ChatGPT to Obtain Search Data and Android Permissions**

原文:Google 被迫开门,ChatGPT 将拿到搜索数据和安卓权限

Summary of Key Points

Regulatory authorities in the United States and Europe are using anti-monopoly measures to target Google's two core strengths in the AI era: its Android operating system and search data. They have demanded that Google grant third-party AI companies (such as ChatGPT and Claude) access to advanced system-level permissions on Android, allowing these AI assistants to directly interact with phone functions just like Gemini can, and to share anonymized search data. These actions are not merely about imposing fines; they aim to break Google's monopoly and create a level playing field for competitors. The implications of these changes will extend to the next round of competition in the AI assistant and search market.

1. The EU Has Issued Two Direct Orders to Google: Open Android to Third Parties and Share Search Data

The EU's approach is straightforward, targeting two critical areas:

  • Granting Super Permissions to AI Assistants on Android: In European Android phones, ChatGPT will be able to use custom wakewords or long-presses on the Home button to access functions directly, read screen content (such as flight and email information), utilize the phone's AI capabilities, and perform actions on behalf of users (e.g., automatically filling in taxi addresses). These features were previously only available with Google's own Gemini. The EU has listed 11 specific functions that must be made available by August 2027 with the Android 18 update. It also requires that any new features added to Gemini be immediately made available to third parties, preventing Google from holding back.
  • Sharing Search Data: Starting from January 2027, Google must share anonymized search data (including ranking signals, user queries, and click records) with qualified competitors (such as Perplexity and local European AI companies). However, this sharing is not free; competitors will need to go through a qualification process and pay a fair fee for the data. Google's search data, which includes billions of user interactions and query modifications each day, represents a valuable resource that competitors would struggle to accumulate in such a short time.

If Google fails to comply, it could face fines equivalent to 10% of its annual global revenue (about $40 billion based on Alphabet's $400 billion in earnings), with the possibility of doubling the penalty for repeated violations, demonstrating the severity of the threat.

2. The US Courts Have Not Broken Up Google, but They Have Exposed Its Core Assets: Sharing Search Data with AI Companies

On the other side of the Atlantic, anti-monopoly proceedings are also underway in the United States. In 2024, Judge Meta ruled that Google had illegally monopolized the search market, but did not break up Chrome or Android. Instead, the court ordered:

  • The sharing of certain search indexes and user interaction data with qualified competitors (including generative AI companies).
  • The provision of a "joint supply service" that allows competitors to use Google's search capabilities for a initial period (up to 40% of the search volume in the first year, gradually decreasing thereafter), providing them with a temporary advantage to catch up technologically.

ChatGPT's CEO stated that the company still needs several years to develop its own technology to handle 80% of user queries on its own, which aligns with the transitional period specified in the US ruling. Although Google has appealed, the regulatory stance is clear: it does not want Google to transfer its search monopoly practices to its AI assistants.

3. Google's Defense of "Privacy and Security": A Strategy Used by Microsoft 20 Years Ago

Google's defense argument is a familiar one: "Sharing data will lead to privacy breaches and threaten national security." Its Chief Legal Officer argued that search records contain sensitive information such as health conditions and financial details, which could be dangerous if shared with unfamiliar companies. However, regulators are not convinced by this rationale. Microsoft used similar arguments when it bundled its IE browser 20 years ago; the real purpose was to maintain its monopoly. While there is a risk of privacy breaches, this cannot be used as a permanent excuse for monopolizing data. Both the EU and US regulations require that data be anonymized before sharing and that only qualified entities are allowed access, aiming to balance privacy and competition.

4. Why Regulators Are Focusing on "Data" and "Access Points"? They Are Targeting Google's Monopoly "Circulation Mechanisms"

Google's monopoly is not solely based on technology but on a self-sustaining cycle:

  • Data Circulation: High search market share leads to more user data, which in turn improves search quality and further increases market share.
  • Access Points: Android defaults to using Google's Gemini as the assistant, creating a habit among users that generates more data and makes Gemini more intelligent.

By targeting these two aspects, regulators aim to break this cycle:

  • Allowing third-party AI assistants (like ChatGPT) direct access to Android without additional steps (e.g., allowing users to book taxis without leaving the email app) reduces user friction; even the best models cannot compete if they require complicated operations.
  • Sharing search data enables competitors to quickly gain the necessary user understanding, eliminating the need for years of data accumulation.

These two aspects represent Google's most valuable assets. Previous anti-monopoly actions only targeted minor issues (such as fines), but this time, the regulators are targeting the fundamental mechanisms of its monopoly.

5. Anti-Monopoly Measures Do Not Help Losers; They Only Give Favors to the Next Winners: Could OpenAI Be the "Lucky One" This Time?

History repeats itself: The 2001 anti-monopoly case against Microsoft did not save Netscape, but it gave rise to Google, a smaller company at the time. Similarly, the current Google case may not revive old competitors like DuckDuckGo, but it could open opportunities for AI companies like OpenAI:

  • The US decision to share search data helps ChatGPT overcome technical shortcomings.
  • The EU's move to open Android access addresses distribution challenges.
  • The five-year transition period specified in the Meta ruling coincides with the time needed for ChatGPT to catch up technologically.

However, there are concerns: OpenAI, now valued at hundreds of billions of dollars and backed by Microsoft, could become the next monopoly. Just as Google benefited from previous anti-monopoly actions, it may also learn to use these regulations to its advantage in the future.

The battle is not over yet. Google will likely argue about details in court (such as the degree of data anonymization and the extent of interface openness), but the overall direction is clear: Android's doors will be opened, and data will be shared. The competitive landscape in the AI era is being reshaped by regulatory authorities.

(The entire analysis is written in plain language, avoiding technical jargon, and each section includes examples and logical explanations to make the complex issues understandable to non-financial readers.)