Apple has unveiled a new framework called Trust Insights in iOS 27, designed to help apps detect when a user may be falling victim to a social engineering scam in progress. The feature runs mostly on-device, analyzing interaction patterns, timing, context, and basic sensor data without inspecting the contents of photos, messages, or emails.
How Trust Insights identifies voice and text scams
According to Apple, social engineering scams are hard to detect automatically because the user often carries out the actions authentically. Trust Insights assigns a medium or high risk level if it detects signs of potential manipulation, allowing the app to trigger warnings, delays, or additional verification steps. The system does not inspect message or call content but evaluates behavioral signals locally and sends only a single output value to Apple servers, where it is combined with account information for a final assessment.
Sponsored Protocol
The five operation categories initially covered by the framework
Trust Insights covers five main categories: tech support scams, authority impersonation, family emergency fraud, voice phishing, and suspicious messages. Apple asks developers to submit feedback via Feedback Assistant for unlisted use cases and to report transactions later confirmed as fraud to improve the system. The feature can be disabled in Settings, but Apple includes a cooldown period to protect users who might have been coached into turning it off.
Sponsored Protocol
For more on digital advertising strategies, read our article on Google Ads Search — Write Ads That Boost Your Quality Score and Lower Costs. Also, learn about proper XML sitemap submission on XML Sitemap in Google Search Console.
The introduction of Trust Insights marks a significant step in protecting users from increasingly sophisticated scams, especially with the rise of AI deepfakes. For technical details, refer to Apple's official documentation on Apple Developer.
