iPhones are about to get a feature that warns users before they are scammed.

Online scams are becoming increasingly sophisticated, especially those that exploit users' psychology to trick them into voluntarily transferring money, providing passwords, or sharing personal information.

In response to this, Apple introduced a new platform called Trust Insights in the new iOS 27 operating system, which helps apps identify when users are at risk of becoming victims of scams in real time.

What is Trust Insights?

Trust Insights is a new application integrated into iOS 27. This platform allows applications to assess the risk level of a transaction or action a user is taking, and then provide appropriate protective measures before damage occurs.

According to Apple, social engineering attacks are one of the biggest challenges today because attackers don't need to attack systems or exploit security vulnerabilities.

Instead, they seek to manipulate users psychologically into performing legitimate actions such as transferring money, sharing verification codes, or changing account information.

Because these actions are performed by the account holder themselves, it is very difficult for traditional security systems to detect and prevent them.

Apple stated that in recent years, scams such as impersonating technical support staff; impersonating authorities or officials; or impersonating relatives calling in emergency situations to request money transfers have increased rapidly.

This risk is further exacerbated by the increasing prevalence of AI-generated audio and deepfake technologies, making it easier for criminals to create realistic-sounding voices or images to deceive victims.

Trust Insights was developed to add a new layer of defense against these hard-to-detect forms of attack.

Detect potential scams without reading message content.

It's worth noting that Trust Insights does not analyze content in Messages, Mail, Photos, or other personal data of users.

Instead, the system primarily operates directly on the device, using behavioral cues to assess risk, including how the user interacts with the device; when actions are performed; usage context; and some basic sensor data on the device.

If Trust Insights detects unusual signs indicating that a user may be being instructed by someone else to commit a scam, it will assess the risk level as medium or high.

This information is then transmitted to the application so that it can proactively implement protective measures.

When Trust Insights assesses an action as potentially fraudulent, the app can take several measures to mitigate the risk, such as displaying a prominent warning to prompt the user to reconsider; temporarily delaying the transaction; requiring identity verification or further confirmation before completion.

The goal is to create an additional "pause" to give users time to think, instead of immediately complying with the scammer's requests.

Apple asserts that it continues to protect privacy.

Apple emphasizes that Trust Insights is designed with privacy as a top priority.

This platform does not read the content of users' emails, messages, or images. The signals are analyzed directly on the iPhone, and the original data is discarded immediately after processing.

Instead of sending all the data to the server, the system only sends a single evaluation result. This value can be combined with some information from the Apple Account and other signs of unusual activity to draw a final conclusion about the risk of fraud.

Users can still turn off the Trust Insights feature.

Apple says users have the option to turn off Trust Insights in Settings if they don't want to use it.

However, the company also revealed that a waiting period may be implemented before the feature is disabled. The purpose is to protect users who are manipulated by fraudsters to disable security layers, making it easier to carry out scams.

In its initial phase, Trust Insights will focus on monitoring five high-risk action groups, including:

- Payment: This includes money transfers, buying and selling digital assets, purchasing goods, and even in-game transactions.

- Account: Change account information, password, or security settings.

- Resource Use: Requests to use resource-intensive or limited services, such as AI inference tasks.

- Communication: Sending messages, submitting electronic forms, or signing documents online.

- Other: Applies to operations that do not fall under the four groups above.

Apple also encourages developers to provide feedback if their app uses the "Other" group, and to report instances that are later confirmed as scams to help Trust Insights continue to learn and improve its accuracy.

The inclusion of Trust Insights in iOS 27 shows that Apple is shifting from a model that only protects the device to one that supports protecting users themselves against increasingly sophisticated scams.

Instead of simply preventing malware or traditional cyberattacks, the new platform aims to identify situations where users may be psychologically manipulated into taking harmful actions.

If implemented effectively, Trust Insights could become a crucial layer of protection, significantly reducing future online scams.