Ant International has introduced AI SHIELD to enhance the security of financial AI for its clients and partners.

To protect AI systems from potential threats and vulnerabilities, AI SHIELD was created. This initiative aims to secure and maintain compliance in AI-driven financial services as AI transforms global finance rapidly.

AI SHIELD is designed to significantly reduce risks related to AI service vulnerabilities, including unauthorized access and data leaks, by up to 90% via advanced assessment methods and proactive safeguards. It employs AI Security Docker to tackle the unreliability of AI services, such as bias, security flaws, and ethical issues.

Details of Ant International’s Launch of AI SHIELD

The official press release indicates that AI Security Docker will ensure secure, reliable, and regulatory-compliant operation when AI is used for fraud detection, payment authorization, and chatbots. It also provides end-to-end protection from the development to deployment stages of AI models and agents to minimize risks.

A significant risk in AI services is over-delegation, which can undermine the reliability of agent-powered payments. Current applications face several major risks when relying on Language Models (LLMs) for payment facilitation due to a lack of strong security controls. Misuse or unauthorized operations by multiple delegated agents are possible without robust safeguards.

AI SHIELD has been developed by Ant International to cut AI service risks by 90% through three main processes: Agent Trustworthiness Authentication (ensuring the security and risk assessment of AI agents before they go live on Ant International’s AI-as-a-Service platform, Alipay+ GenAI Cockpit), AI Service Safeguard (offering real-time monitoring and intervention to block emerging threats via multi-layer intent recognition), and Dynamic Patrolling & AI Red Teaming (ensuring continuous inspection and adversarial testing for swift resolution of vulnerabilities).

Furthermore, AI-powered risk management tools safeguarded by AI SHIELD will protect transactions throughout their lifecycle at Ant International. These solutions are intended to shield companies, businesses, and users from deepfake attacks, fraud, and scams.