Meta rolls out new AI tools to spot scams across its apps

In short: Meta announced new AI-based scam detection tools for WhatsApp, Facebook, Messenger, and Instagram.

What happened

Meta said it is rolling out a set of automated systems to find and stop scams across its main apps. The company says these tools look at many clues at once, including the words in messages and posts, images, account behavior, location patterns, and other context. Think of it like a security guard who does not just read the sign on a package, but also checks the return address, the packaging, and the delivery route.

Meta said the goal is to catch more kinds of scams, including fake accounts pretending to be a celebrity or a brand, links that send people to harmful sites, and tricks that copy real website addresses. It also highlighted scams tied to linking a device to WhatsApp using a QR code, which can be abused to take over an account.

Meta pointed to 2025 numbers to show impact. It said it removed 159 million scam ads globally, and claimed about 92 to 93 percent were caught before users reported them. Meta also said it disabled 10.9 million accounts tied to fraud, and highlighted an operation in early 2026 where it removed more than 150,000 accounts linked to scam centers in Southeast Asia.

Why it matters

Scams often move fast, especially when they follow big news events. Meta says these systems can respond quicker and make fewer mistaken takedowns, but it has not shared independent test results that show how often the AI is right or wrong.

Source: TechCrunch AI