日本語 
Cognizant has launched Cognizant Secure AI Services, a new integrated service designed to help businesses securely protect, govern, and scale AI and agent-based systems across their entire operations.
As AI systems are deployed company-wide, organizations are increasingly integrating AI into decision-making, automation, customer engagement, and key workflows. These systems are becoming agent-based, capable of autonomously performing reasoning, dialogue, and task execution while interacting with enterprise data, various systems, and external applications. While this shift holds the potential to transform businesses, it also introduces new security, governance, and runtime risks that traditional cybersecurity models cannot adequately address.
Traditional security has been built on the premise of deterministic software whose behavior is predetermined. On the other hand, AI systems operate based on probability and adapt to the situation, making them vulnerable to exploitation in ways that traditional security tools cannot detect. Tampered models, malicious prompts, and agent malfunctions can lead to widespread misbehavior.
Also Read: Serverworks Introduces Wiz Security Platform Support
The services announced today are designed to help companies move from “tacit trust” to “provable trust” based on evidence, traceability, and ongoing assurance. Cognizant ensures reliability for AI systems through a two-stage approach: build and operate. First, in the build phase, models, data, and pipelines are protected before deployment. Second, in the operate phase, AI behavior is continuously monitored in the production environment to help detect malicious activity, manage and mitigate security vulnerabilities, and store evidence to support audits.
“Cognizant Secure AI Services” is built on the following three foundations:
“Agent Development Lifecycle (ADLC)”
Protection features are built in at every stage of the AI system’s design, construction, testing, deployment, and modification.
“Cognizant Neuro® Cybersecurity”
This system centrally manages AI and signals across the entire enterprise system for threat response, correlation analysis, and evidence preservation for audits.
“Responsible AI”
This is a continuous layer of trust and assurance provided through “Cognizant Trust™”. As AI systems scale, it supports traceability, policy enforcement, and compliance based on customer-defined requirements.
By working together, these functions comprehensively cover aspects such as protecting AI models and data, managing user permissions, ensuring the secure operation of AI systems, managing the behavior of AI agents, and mitigating risks when using generative AI. The goal is to enable companies to securely manage their systems at every stage of AI operation.
Cognizant already works with more than 250 companies worldwide, primarily in highly regulated industries, to evaluate, protect, and operate digital transformation programs, including AI deployments. Early implementations address some of the most critical risks facing organizations today, from preventing deepfake fraud and AI model tampering to protecting autonomous agents and generative AI systems running across enterprise workflows. At the same time, we are working with clients to build the governance and audit frameworks necessary to responsibly scale AI in a regulated environment.
SOURCE: PRTimes
English