[Digital Today reporter Chi-gyu Hwang] French defence and security company Thales is expanding cooperation with Google Cloud, Techzine reported on Sept. 29 local time.

According to the report, Thales will combine its "AI Security Fabric" with Google Cloud's "Gemini Enterprise Agent Platform". This will allow companies to protect, manage and monitor in real time interactions flowing between users, agents, models and tools.

The partnership is focused on addressing concerns that arise as companies deploy AI agents in real work.

Agents differ from existing applications that move through fixed procedures. They reason and plan on their own to make decisions. They also communicate with other agents.

Thales explained that this creates new attack paths that are difficult to block with existing security architectures. It cited prompt injection, sensitive data leakage, unsafe outputs, unauthorised actions and increasingly complex agent-to-agent interactions as risk factors.

AI Security Fabric implements controls around agents so they operate only within allowed limits and restricts data access. It blocks inappropriate actions in real time. Companies can use it to apply policies, detect threats targeting AI and continuously monitor agent behaviour. It also provides centralised governance and regulatory compliance functions.

Eva Roudin (에바 루딘), senior vice president of Thales' cybersecurity products division, said, "Companies are moving from AI assistants to AI agents that act and decide on their own and communicate with core business systems," and added, "Security also needs a fundamentally different approach."

Vinit Bhan (비닛 반), head of security and identity partnerships at Google Cloud, said, "As agents are connected to core data and systems, customers want an environment with security and governance built in," and added, "Cooperation with Thales will add capabilities to protect agent AI work."