Meta Launches New AI Security Tools and Meta AI App

What happened

Meta launched several new AI security and privacy tools alongside its Meta AI app on April 29, 2025, at the LlamaCon event. These include LlamaFirewall to prevent malicious AI activities, CyberSec Eval 4 for cybersecurity benchmarking, and the Llama Defenders Program to support partner organizations in AI security. The Meta AI app is powered by Llama 4 and integrates with Facebook, Instagram, and Meta AI glasses.

Why it matters

• These tools aim to enhance the security and privacy of open-source AI applications.
• The Llama Defenders Program helps organizations access advanced AI security solutions.
• Meta’s AI app offers personalized AI interactions integrated with social media and wearable devices.

Key facts

1. 1. Meta AI app and security tools were launched on April 29, 2025, during LlamaCon.
2. 2. LlamaFirewall is a new tool designed to prevent malicious activities targeting AI models.
3. 3. CyberSec Eval 4 includes two new tools, CyberSOC Eval and AutoPatchBench, for assessing AI defense capabilities.