Google Confirms Android SafetyCore Enables AI-Powered On-Device Content Classification

Google Confirms Android SafetyCore Enables AI-Powered On-Device Content Classification
Summary: Google has clarified that its newly introduced SafetyCore app for Android does not engage in client-side scanning of user content. This system service is designed to help users detect unwanted content while ensuring user privacy and control over their data. Developed as part of a security initiative, SafetyCore utilizes on-device machine learning to categorize content when requested by applications, without attempting to identify illegal material.

Affected: Google Android System and users of Android 9+ devices

Keypoints :

  • SafetyCore provides on-device machine learning capabilities for classifying content as spam, scam, or malware.
  • The service is designed to prioritize user privacy and requires consent from users to operate.
  • It contrasts with client-side scanning methods, which have raised significant privacy concerns.

Source: https://thehackernews.com/2025/02/google-confirms-android-safetycore.html