Customer Cloud Corp. has unveiled “CC AGI Security,” a new security initiative to ensure the safety of enterprise AI platforms and digital services in the AI era. With the spread of generative AI, a growing number of companies are building their own dedicated AI platforms (local LLMs) and developing and operating AI services and AI products. At the same time, AI systems are increasingly linked with existing web services and cloud systems, bringing to light security challenges that differ from those of the past.
Against this backdrop, we have unveiled the concept of “CC AGI Security,” a new security domain covering AI platforms and digital services as a whole, and are currently advancing research and detailed design.
Security challenges in the AI era
With the spread of generative AI and AI agents, new kinds of risk are emerging in corporate IT environments. Examples include AI-specific attacks such as prompt injection, unauthorized access to AI models and training data, data poisoning, the malfunction or misuse of AI agents, and new types of cyberattacks targeting AI systems.
In addition, AI applications are often operated in conjunction with web services and cloud systems, increasing the importance of security measures that protect AI platforms and existing systems as a single whole.
Such risks cannot always be adequately addressed by conventional checklist-based or perimeter-based security. With the arrival of AI, the speed, scale, and sophistication of attacks are changing all at once, and we have reached a point where the very premises of security need to be reconsidered.
From an era of “using” AI to an era of “owning” it.
The domains to be protected must also be redesigned from the ground up.
Overview of CC AGI Security
CC AGI Security is a security domain covering AI platforms and digital services as a whole. Its aim is to ensure the safety of a wide range of digital services — not only the AI platforms (local LLMs) and AI applications that companies build, but also general web services and cloud systems.
The main areas covered are as follows.
- AI platform security — Safety verification of dedicated enterprise AI platforms (local LLMs) and AI models
- Safety assessment of AI applications — Behavior verification of AI agents and AI services
- Security of web services and cloud systems — Safety verification of general web applications and cloud environments
By addressing new architectures in which AI and existing systems are integrated, we will provide the security measures required of digital services in the AI era.
Our work to date
Customer Cloud has previously announced initiatives such as “CC LLM,” which builds a dedicated AI platform for each company, and “CC AI Factory,” a model for continuously developing AI services.
CC AGI Security is an initiative that anticipates the spread of such AI platforms and establishes a security domain covering AI systems and digital services as a whole. Based on the philosophy of “Non-Destructive Innovation” — layering what the AI era requires on the outside of existing security structures rather than replacing them — we will establish a new kind of safety for the AI era while respecting the operations and assets that companies have built up.
Next steps
Going forward, we will pursue the following initiatives.
- Safety assessment of AI systems
- AI security research
- Strengthening the security of AI development environments
- Hosting AI security seminars
As generative AI spreads as a foundation of society and industry, ensuring the safety of AI systems has become an important issue. Customer Cloud will continue to build out a security domain suited to the AI era.