Apple is taking a major step toward combating social engineering attacks with a new feature in iOS 27 that can warn users in real time if they are likely being targeted by a scam.
The new framework, called Trust Insights, is designed to detect suspicious behavioral patterns across user interactions, including apps, calls, messages, and other activities.
Unlike traditional security tools that scan for malicious files or links, Trust Insights focuses on identifying behavioral signals indicating that a user may be manipulated into taking risky actions.
Apple highlights that modern scams often rely on psychological tactics rather than technical exploits, making them harder to detect using conventional methods.
The Trust Insights framework runs on-device primarily and analyzes factors such as interaction timing, contextual behavior, and basic sensor data.
iPhone Scam Alerts
This allows the system to detect if a user is being guided or “coached” through actions commonly associated with scams, such as transferring money, sharing sensitive information, or changing account settings.
According to 9to5Mac, Trust Insights rates suspicious activity as medium or high risk, enabling apps to warn users, delay actions, or request additional verification before proceeding.
This approach aims to disrupt ongoing scams without significantly disrupting legitimate user activity. Apple emphasized privacy as a core component of the framework. Trust Insights does not inspect the content of Messages, Mail, Photos, or other personal data.
Instead, it processes behavioral data locally on the device and discards raw data immediately after analysis. Only a single risk signal is transmitted to Apple’s servers, where it may be combined with account-level indicators such as unusual login behavior to produce a final assessment.
The framework supports five key operation categories where scams are most likely to occur:
- Payments: Financial transactions and in-app purchases.
- Account Changes: Credential and security setting updates.
- Resource Usage: High-cost actions, such as AI processing.
- Communication: Sending messages or signing documents.
- Other: Any remaining sensitive activities.
Apple has also built safeguards into the system to prevent attackers from disabling protections.
While users can turn off Trust Insights in device settings, Apple notes there may be a cooldown period before changes take effect. This is intended to protect users who may have been coerced into disabling security features.
Developers are encouraged to integrate Trust Insights into their apps and provide feedback to Apple at the WWDC26 Conference. This includes reporting how the framework impacts user transactions and flagging confirmed fraud cases to improve detection accuracy over time.
The introduction of Trust Insights comes amid a sharp rise in social engineering attacks, including tech support scams, impersonation fraud, and AI-driven deepfake schemes.
By focusing on real-time behavioral analysis and privacy-preserving machine learning, Apple aims to provide a proactive defense layer against evolving threats.
With iOS 27, iPhone users may soon receive timely warnings to help them avoid scams, marking a shift from reactive security to real-time intervention.
