Ant Financial's continuous innovation in AI security is becoming a leading security practice in the industry, ensuring that customers can enjoy the benefits brought by AI intelligent agents while still having controllable and traceable security. "AI intelligent agents are not 'black boxes,' much less 'blind boxes.'"
Article author and source: AIBase
As AI agent technology rapidly penetrates core business scenarios, autonomous AI, executor-based like OpenClaw, is becoming a key engine for a new round of productivity revolution. However, its powerful system access and task execution capabilities also bring unprecedented security risks—problems such as privilege escalation, memory corruption, malicious manipulation, and computing power abuse are becoming increasingly prominent. While embracing efficiency, some enterprises have underestimated the potential governance challenges.
To this end, Ant Financial officially launched the "Ant Sky Detection 2.0 - Lobster Guardian" AI security protection system and simultaneously launched the "Lobster AI Security Protection Plan," aiming to provide enterprises that deploy OpenClaw with full lifecycle and multi-level security protection.
The newly released "Claw Security Suite 1.0" officially launched on March 19th, focusing on three core capabilities: First, "Combating Ideological Mutation," which accurately identifies input and output content, intercepts sources of illegal information and malicious inducement, and prevents models from being "blackened"; second, "Purifying the Skills Repository," which, based on Ant Financial's CALIR five-dimensional model framework, performs compliance scanning and behavior auditing on Skills plugins, effectively identifying and blocking risks such as overreach of permissions, sensitive information disclosure, and abnormal calls; and third, "Risk Sentiment Broadcasting," which generates real-time risk emergency reports and proactively pushes them to help enterprises grasp the security situation at the first moment, achieve transparency and immediate perception of risks, and transform passive defense into proactive response.
To help more enterprises complete the construction of security baselines at low cost, Ant Financial has launched the "Lobster AI Security Guardian Program". The program will provide the first batch of 100 partner companies with 200,000 free calls to Lobster security protection services over a period of 3 months, aiming to allow customers to quickly experience enterprise-level AI security protection.
In the field of large-scale security technology, Ant Financial has been successfully included in IDC's "China Large-Scale Security Market Map" thanks to its leading technological strength and solutions. Based on real-world attack and defense experience, Ant Financial's large-scale security technology has achieved state-of-the-art (SOTA) performance in complex scenarios such as financial compliance and jailbreak attacks, and has the ability to generalize and repair risks from 0-day.
Ant Financial's continuous innovation in AI security is becoming a leading security practice in the industry, ensuring that customers can enjoy the benefits of AI agents while still having controllable and traceable security. "AI agents are not 'black boxes,' much less 'blind boxes,'" said the Ant Financial AI security team. "We hope that through 'Lobster Guardian,' enterprises can enjoy the benefits of AI agents while having an explainable, controllable, and traceable security foundation."
