Ant International has announced the launch of its AI SHIELD to improve financial AI security for clients and partners.
Following this announcement, AI SHIELD was developed in order to protect AI systems from threats and vulnerabilities, as well as to optimise the way AI-powered financial services stay secure and compliant as AI rapidly transforms global finance.
In addition, AI SHIELD is set to reduce the risk of AI service vulnerabilities, such as unauthorised access and data leakage by 90% through the use of improved assessment methods and running safeguards. AI SHIELD uses AI Security Docker in order to address the unreliability inherent in AI services, including bias, security vulnerabilities, and ethical concerns.
More information on Ant International’s launch of AI SHIELD
According to the official press release, the AI Security Docker is expected to make sure that AI, when leveraged for things like fraud detection, payment authorisation, and chatbots, operates securely, reliably, and in compliance with regulations. At the same time, it will also provide end-to-end protection from development to deployment of AI models and agents in order to reduce AI service risks.
A common AI service risk is over-delegation, which undermines the overall reliability of using agents to facilitate payments. Through the strategy of using agents to process payments to offer new levels of automation and efficiency in financial services, current agent applications face several major risks when relying on LLMs. Without robust security controls, multiple agent-delegated services could be misused or perform unauthorised, unsafe operations.
Ant International’s AI Security Docker was developed to reduce AI service risks by 90% through the following three key processes. These include the Agent Trustworthiness Authentication (a security evaluation and risk discovery for AI agents before going live on Ant International’s AI-as-a-Service platform Alipay+ GenAI Cockpit), the AI Service Safeguard (which provides real-time monitoring and intervention on agent interactions in order to block emerging threats leveraging multilayer intent recognition), and the Dynamic Patrolling & AI Red Teaming (representing an ongoing inspection and adversarial testing that ensures rapid resolution and reinforcement against vulnerabilities)
In addition, AI-powered risk-management tools secured by AI SHIELD will provide full-lifecycle protection of transactions at Ant International. These solutions are expected to protect companies, businesses, and users from deepfake attacks, fraud, and scams.
