Meta expands anti-scam tools across WhatsApp, Facebook, and Messenger with more AI detection

What Meta announced

On March 11, 2026, Meta announced a broad anti-scam push across its consumer apps, combining new user-facing warnings with expanded AI-based detection and stronger advertiser verification. The scope matters. This is not one feature inside one product. Meta is updating protections across WhatsApp, Facebook, and Messenger, while also tightening the ad-side controls that scammers can exploit to reach people at scale.

Several of the changes are designed to surface risk before users engage. Meta said it is strengthening device-linking awareness in WhatsApp, testing suspicious friend request alerts on Facebook, and expanding advanced scam detection on Messenger to more countries. In Messenger, when a chat with a new contact matches patterns associated with common scams such as suspicious job offers, users can choose to share recent messages for an AI scam review. If the system flags a likely scam, Meta provides information on common tactics and suggests actions such as blocking or reporting the account.

How Meta says the system works

• Meta said its AI systems analyze text, images, and surrounding context to detect more sophisticated scam patterns.
• The company is using AI to catch celebrity, public figure, and brand impersonation, including deceptive fan-style framing, misleading bios, and domain mimicry.
• Meta said it is expanding advertiser verification so that verified advertisers account for 90% of ad revenue by the end of 2026, up from 70% today.
• The company also disclosed that it removed more than 159 million scam ads in 2025, with 92% taken down proactively before user reports.

Why it matters

The release reflects a broader shift in online fraud defense. As scammers use generative systems to scale deception, platform operators are responding with more context-aware AI that can evaluate images, language, account behavior, and links together rather than through static rules. Meta is effectively saying that modern scam detection requires a model-driven trust and safety stack, not just a moderation queue.

It also shows that fraud prevention is becoming a cross-surface infrastructure problem. User warnings, advertiser identity checks, proactive ad enforcement, and impersonation detection all need to work together because scam operations move across messaging, ads, social profiles, and external websites. Meta's update is significant not only because of the individual tools, but because it packages scam defense as an end-to-end platform capability rather than a set of isolated safety features.

Source: Meta official announcement