AI agents going out of control will trigger a 'digital mega fire'… What solutions does Rubric propose?

As companies introduce artificial intelligence ( AI ) to build autonomous working environments, the control permissions and security management of AI agents are becoming core issues in digital operations. In particular, highly developed AI agents can perform tasks or make business decisions on behalf of user identification, and once security is breached, it could lead to the worst-case scenario of the entire organizational system being taken over.

Bipul Sinha, co-founder and CEO of Rubrik, emphasized in a recent interview: “In an era where agents handle business at superhuman speed, damage can occur tenfold in just ten minutes.” He pointed out that the traditional security system struggles to manage 'agent governance', which is emerging as a new means of control. He explained, “AI agents operate business processes like humans and sometimes even mimic user identification, thus possessing the attributes of 'super humans'.”

To enhance agent security, Rubrik has launched a new platform called 'Rubrik Agent Cloud'. This system can analyze the types and activity scope of AI agents within the organization, detect threat behaviors based on preset protective barriers, and provide the functionality to quickly withdraw the agent when necessary. CEO Sinha explained: “All AI agents should be treated like new employees, equipped with a code of conduct manual, monitoring system, and a pause procedure when issues arise.”

But the real challenge of agent governance lies in the speed of large-scale diffusion. Allowing agents to spread uncontrollably in the absence of sufficient visibility and control may lead to small-scale failures evolving into a ‘digital major fire’ that paralyzes the entire system. In this regard, CEO Sinha suggests that organizations should not formulate overly large automation plans, but should limit them to 3-4 core workflows, completing pilot operations before gradually expanding the scope.

Sin Ha also warned: “The operating environment must use precise AI, and if there is no cost management capability, the reasoning costs will spiral out of control.” This means that the reality of cost stability cannot be ignored in the AI agent diffusion strategy.

In the current context where AI technology is deeply integrated into the overall enterprise, Rubrik's proposed 'agent governance' strategy has transcended mere security measures and is becoming the core control tower for the concrete realization of organizational overall digital resilience and sustainability.