Apple is introducing the Trust Insights framework in iOS 27 to help applications detect and mitigate social engineering scams by analyzing user behavioral patterns and interaction context on-device.

Key Points

• The framework monitors interaction patterns, timing, and sensor data to identify potential fraud without accessing private content in Photos, Messages, or Mail.
• Trust Insights assigns risk levels to transactions, allowing apps to trigger warnings, delays, or additional verification steps for users.
• The system covers five primary operation categories: payment, account management, resource use, communication, and a fallback category for other activities.
• Processing occurs primarily on-device, with only a single output value sent to Apple servers to maintain user privacy.
• Users can disable the feature in settings, though a cooldown period is enforced to prevent scammers from coercing victims into turning off protections.

Why it Matters

This framework represents a significant shift in mobile security by addressing the human element of fraud that traditional automated filters often miss. By providing developers with real-time risk assessments, Apple aims to curb the rising prevalence of AI-driven scams and authority impersonation attacks.