Global Scam IntelligenceDaily updates on global scam news and tactics
Back to home
PhishingPublished: August 16, 2026🌐

WhatsApp Tests New 'Scam Alert' Feature Using On-Device AI to Detect Fraudulent Messages

WhatsApp has begun limited beta testing of a new 'Scam Alert' feature in mid-August that uses artificial intelligence to identify fraudulent messages from unknown contacts directly on users' devices. The technology maintains end-to-end encryption by analyzing messages locally rather than on external servers, protecting user privacy while detecting common scam patterns.

WhatsApp's New 'Scam Alert' Feature Explained

Overview of the Feature

WhatsApp launched limited beta testing of 'Scam Alert' in mid-August 2025, a feature designed to detect fraudulent messages from unknown senders using artificial intelligence. Crucially, all analysis occurs locally on the user's device, ensuring that end-to-end encryption remains fully intact. No messages are sent to external servers, preserving complete privacy.

How It Works

The feature uses an on-device machine learning model that examines incoming messages from contacts not stored in the user's address book. The system identifies typical fraud patterns, including romance scams and attempts to create artificial urgency—common tactics used by scammers. When suspicious activity is detected, a warning banner appears on the recipient's screen visible only to them.

User Options When Alerted

When a warning appears, users have several options:

  • Block the sender
  • Report the message
  • Mark the contact as trustworthy

Once a contact is marked as safe, no further warnings appear for that conversation.

Privacy Protection Measures

Meta emphasizes that message contents are never transmitted to the company, protecting user privacy. However, users have the optional choice to send the last five messages of a reported chat to Meta to help improve the AI model. Additionally, model versions are made publicly available through a transparency ledger accessible via Cloudflare, and aggregated data collection uses Differential Privacy techniques to prevent individual identification.

The Fraud Problem and Context

This feature addresses a critical security issue. According to the US FTC (Federal Trade Commission), fraud through social media caused $2.1 billion in total losses in 2025, with WhatsApp-related fraud alone accounting for $425 million of that amount.

In 2025, Meta removed 159 million scam ads from its platforms, with 92 percent removed before users reported them.

Future Rollout

No specific launch date has been announced for widespread deployment. Meta is currently conducting limited testing to validate the effectiveness of detection patterns in real-world conditions.

How to Protect Yourself

  • Be cautious of messages from unknown contacts: Messages from people not in your address book deserve careful consideration, especially unsolicited ones.
  • Question messages creating urgency: Scammers often pressure victims into quick action. Be skeptical of messages demanding immediate responses or personal information.
  • Use the Scam Alert feature: When a warning banner appears, trust it. Block or report suspicious contacts rather than ignoring the alert.
  • Combine with device security: Use additional security software on your device alongside WhatsApp's built-in protections.
  • Never share sensitive information: Do not share passwords, financial details, or verification codes with unknown contacts.

Where to Report Fraud

If you suspect fraud or have been victimized, report it immediately. Use WhatsApp's built-in reporting function and contact local law enforcement or consumer protection agencies in your country. Timely reporting helps authorities track scam patterns and protect others.

Source: ad-hoc-news.de

Share