Meta is deploying advanced AI tools and user alerts across Facebook, WhatsApp, and Messenger to proactively block scams. This focused strategy aims to automate the detection of scams and protect users from digital threats, including impersonation, fake investment schemes, and account takeovers.
Key Anti-Scam AI Tools and Features (March 2026)
- AI-driven systems on Facebook and Instagram each analyze text, images, and context to spot scams. On Facebook, AI identifies fake celebrity pages, brand impersonations, and fake fan accounts. On Instagram, it targets similar scams and influencer impersonations. These systems also scan for suspicious connections, misleading wires, and links mimicking real websites, tailoring strategies to each platform’s common scam types.
- Meta is testing a Facebook-specific alert that actively warns users about suspicious friend requests. Alerts trigger when requests come from accounts with few mutual friends, mismatched locations, or recently created profiles. This AI-driven Facebook tool helps users recognize scam attempts unique to friend connections.
- To address account takeover scams, WhatsApp’s new AI tool warns users about suspicious device-linking requests. For example, if a fake QR code or a phishing voting scheme initiates a linking attempt, WhatsApp’s AI highlights the request’s suspicious origin to help prevent scams unique to its platform.
- Messenger is expanding its AI feature to more countries, allowing it to scan new chats for scam patterns such as job scams and fake investment schemes. Messenger’s AI sends warnings and lets users submit chats for AI safety review, focusing on suspicious conversational scams common to this platform.
- In 2025, Meta removed more than 159 million scam ads. AI detected 92% of these before users reported them.
- In 2025, Meta deleted 10.9 million Facebook and Instagram accounts linked to scams.
Long-term Strategy and Partnerships
- By the end of 2026, Meta intends for 90% of its ad revenue to come from verified advertisers, which is expected to bolster the effectiveness and integrity of platform ads as part of the anti-scam strategy.
- Meta is collaborating with banks and law enforcement to dismantle criminal networks, leading to the deactivation of over 150,000 accounts linked to scam centers in Southeast Asia.
These AI tools are the cornerstone of Meta’s strategy to automate content moderation, improve operational efficiency, and minimize reliance on external vendors.
Meta says its AI enforcement technology outperforms human review teams in finding fake accounts and sexual solicitation content.
The company announced on its website Thursday that it was rolling out the Meta AI Support Assistant globally on Facebook and Instagram. This tool will provide 24/7 support for account issues, including password changes and profile settings. It will also revamp the company’s approach to content enforcement, making it more effective at identifying and removing severe violations, such as scams and illegal content.
Meta said the AI expansion will occur over the next few years.
Meta stated we are launching new AI enforcement and support tools to strengthen safety and user experience across our apps. As technology advances, these AI capabilities will provide faster, more reliable, and more consistent detection of serious violations like this. Scones
Launching the Meta AI Support Assistant
Meta previewed its AI Support Assistant in December. It is now launching the Assistant in places where Meta AI is available on Facebook and Instagram apps for iOS and Android. It is also available in the Help Center.
The Meta AI Support Assistant addresses account issues and responds to questions about notification settings or new features. It also offers support for:
- Reports of Scams, Impersonation Accounts, or Problematic Content
- Questions about why the content was taken down and how to appeal these decisions.
- Handling Privacy Settings.
- Resetting passwords.
- Updating Profile Settings
The Meta AI Support Assistant is built into both Facebook and Instagram, providing rapid responses for account-related queries on both apps, typically within 5 seconds.
Meta described the AI Support Assistant as an important step toward delivering stronger support within its applications.
The assistant is being launched in all languages that Facebook and Instagram support for health topics.
Improving Content Enforcement
Meta faces criticism for easing moderation but says it remains focused on reducing mistakes and proactively targeting the most severe illegal content, including terrorism, child exploitation, drugs, fraud, and scams.
Meta says it is testing advanced AI systems for content enforcement. These systems can catch more violations accurately, stop more scams, and respond faster to real-world events with fewer over-enforcement mistakes.
Meta said its new AI systems can:
- New AI systems can reduce the likelihood that scammers trick people into giving up their login details. These systems now detect and handle 5,000 scam attempts per day that no existing review team had previously caught.
- Identified and prevented more accounts from impersonating celebrities and other high-profile people, which helped us reduce user reports of the most-impersonated celebrities by over 80%.
- The AI can now catch twice as much violating adult sexual solicitation content as review teams. It reduces the error rate by more than 60%.
- The AI can prevent account takeovers by noticing certain behaviors. For example, it monitors whether an account is accessed from a new location, whether the password is changed, or whether profile edits are made. While these changes might seem harmless to a person, the AI can identify them as a threat.
- Detect a Fake Site Spoofing a Legitimate Website. The AI can detect fake sites pretending to be real stores by spotting things like a real logo used with very low prices and a suspicious web address. This is in languages spoken by 98% of people online, far beyond our previous coverage of around 80 languages, according to Meta.
More Advanced AI Systems
Over the next few years, Meta will deploy these advanced AI systems across its apps. Deployment will begin once they consistently outperform current content enforcement methods. This shift will change how the company handles enforcement.
Meta plans to rely less on third-party vendors for content enforcement. The company will focus on building up its own systems and staff.
While content review by people will continue, these AI systems will address tasks suited to technology, such as repetitive graphic content review or evolving challenges posed by illegal drug sales and scams.
AI can help us move faster and operate at scale, but it doesn’t replace human decision-making. It helps us apply it more consistently across billions of pieces of content on our platforms, the company said in its announcement. Experts will design, train, oversee, and evaluate our AI systems, measuring performance and making the most complex, high-impact decisions. For example, people will continue to play a key role in how we make the highest-risk and most critical decisions, such as appeals of account disablement or reports to law enforcement.
Meta also pledged that its community standards won’t change as part of the shift to AI, and that it will improve its methods for reporting, handling violations, and addressing mistakes.
Source: https://www.sanjoseinside.com/news/meta-reveals-plan-to-gradually-replace-human-moderators-with-ai/
