WhatsApp has started testing a new security feature to help users identify possible scam attempts received by the app. The tool uses artificial intelligence to analyze suspicious messages and issue an alert within the conversation itself.
WhatsApp tests AI alert to identify possible scams
WhatsApp has started testing a new security feature to help users identify possible scam attempts received by the app. The tool uses artificial intelligence to analyze suspicious messages and issue an alert within the conversation itself....
This means that the content of messages does not need to be sent to WhatsApp servers for the system to classify. When the artificial intelligence model identifies signs of possible fraud, a warning appears in the conversation.
- The person on the other side cannot see this alert.
- From there, the user can choose how to proceed.
Editorial reading aid based only on information contained in this story and its identified source.
According to the company, processing takes place directly on the user's device. This means that the content of messages does not need to be sent to WhatsApp servers for the system to classify.
When the artificial intelligence model identifies signs of possible fraud, a warning appears in the conversation. The person on the other side cannot see this alert.
From there, the user can choose how to proceed. Options include blocking the contact, reporting the conversation or continuing to exchange messages normally.
User can indicate when alert is wrong
WhatsApp also foresees situations in which artificial intelligence may misinterpret a message. If this happens, it will be possible to classify that conversation as trustworthy.
“If the model identifies a message as a likely scam attempt, the user sees a warning in the conversation, which is not visible to the other person,” explains the company.
When you mark a conversation as safe, the alert disappears and will not appear again in that chat.
The company also intends to use feedback from users to improve the tool's accuracy. Anyone who considers the conversation to be trustworthy can, if they wish, share the last five messages received with WhatsApp.
This sending, however, will not be automatic.
“If a user marks a conversation as trustworthy, they have the option to share the last five messages received with WhatsApp to help improve the accuracy of the functionality,” the platform said.
WhatsApp says messages remain private
When presenting the new system, WhatsApp reinforced that the analysis carried out by artificial intelligence occurs locally on the cell phone.
According to the company, “none of the message content leaves the user’s device for classification.” Sending excerpts of a conversation to WhatsApp will only happen if the person themselves decides to share this information to help improve the feature.
The tool will also be optional. Anyone who has access can deactivate it if they do not want to receive warnings about possible scams.
Feature is still in the testing phase
Alerts have not yet been released to all WhatsApp users. At the moment, the new feature is in a limited beta version, used to test the functionality and improve the system before an eventual wider distribution.
The initiative seeks to reinforce the application's protection tools in a scenario in which messages sent by unknown numbers are often used in fraud attempts and other types of digital scams.
WhatsApp will stop working on cell phones with Android 5
The application will require Android 6 or later from September 8th and will raise the minimum requirement for iPhones to iOS 15.5 on November 30th. Users of old devices should check the installed version and possible updates