Image Courtesy: Digital Trends
Apple is introducing a new security feature in iOS 27 designed to help apps recognize when users may be falling victim to social engineering scams as they happen. The new framework, called Trust Insights, analyzes behavioral patterns in real time and allows apps to intervene before sensitive actions are completed.
The company says Trust Insights is built to address scams that rely on manipulating users into authorizing payments, sharing personal information, or changing account settings themselves, making them much harder to detect through traditional security systems. Rather than examining the content of messages, emails, or photos, the framework evaluates interaction patterns, timing, context, and limited sensor data, most of which is processed directly on the device.
If the system detects behavior consistent with a user being coached through a scam, it assigns either a medium or high risk rating. Apps can then respond by displaying warnings, delaying high-risk actions, or requiring additional verification before completing transactions.
Apple emphasized that Trust Insights is designed with privacy in mind. The framework does not inspect the contents of Messages, Mail, or Photos. Instead, it immediately discards the behavioral data after processing it on the device and sends only a single risk value to Apple’s servers. That value may then be combined with Apple Account activity and other security signals before a final assessment is returned.
Users will have the option to disable Trust Insights in Settings, although Apple says a temporary cooldown period may apply before the change takes effect. The company says this safeguard is intended to protect users who may have been manipulated into turning off the feature during an active scam.
At launch, Trust Insights will support five categories of actions, including payments, account changes, resource-intensive requests such as AI processing, communications like sending messages or signing documents, and a general category for other operations. Apple is also asking developers to report how well the framework performs and identify confirmed fraud cases to help improve future scam detection.
