Google Chrome's version 148 update silently removed a phrase from the on-device AI settings that previously assured users data would not be sent to Google servers. In Chrome 147, the description stated that AI models run directly on the device "without sending your data to Google servers." This line is now absent in Chrome 148.0.7778.97, replaced by a simpler statement that AI models run on the device, with no mention of data handling. The change was first spotted on the Chrome subreddit and quickly gained attention on Hacker News, sparking privacy concerns among users. The removal comes amid ongoing controversy over Chrome's silent download of a 4GB Gemini Nano AI model file without user consent. Privacy researcher Alexander Hanff confirmed the behavior across platforms and argued it violates EU privacy laws. Google's original justification for the silent install was that on-device processing kept data off its servers, a claim now undermined by the text removal. For wallet and key holders using Chrome, this development raises concerns about potential data leakage through browser-based AI features. Even if data is processed on-device, any interaction with websites using the AI model could expose inputs and outputs to those sites under their privacy policies. Users should be cautious about using Chrome's AI features for sensitive tasks, especially those involving cryptocurrency, private keys, or personal data, as the current privacy protections may be less than advertised.

关键事实

• Chrome 148 removed phrase that said data not sent to Google servers.
• Phrase was in on-device AI settings; change spotted on Reddit and Hacker News.
• Chrome silently downloads 4GB Gemini Nano model with no opt-in prompt.
• Privacy researcher argues download violates EU ePrivacy Directive Article 5(3).
• Google says data still processed on-device, but websites may see AI inputs/outputs.

KeyAudit 数据视角