As companies rapidly adopt agentic AI systems capable of acting autonomously and making decisions with limited human oversight, cybersecurity leaders are warning that trust, identity management, and governance will become critical challenges. In a recent discussion, Cisco executive Tom Gillis emphasized that businesses cannot safely scale autonomous AI agents without building strong frameworks for authentication, access control, and accountability. Experts argue that as AI systems begin interacting with networks, applications, and sensitive corporate data independently, traditional security models may no longer be sufficient.
One major concern involves identity verification for AI agents themselves. Unlike traditional software tools, agentic AI systems can make decisions, initiate actions, and communicate with other systems autonomously. Security researchers warn that organizations must establish mechanisms to verify which AI agents are trusted, what permissions they have, and how their activities are monitored. Cisco and other cybersecurity firms increasingly advocate for “zero trust” architectures and microsegmentation strategies that isolate systems and restrict AI agents to tightly controlled environments to limit damage if something goes wrong.
The urgency around AI governance has intensified as evidence grows that cybercriminals are already using AI to accelerate hacking operations. Google recently reported one of the first known cases where hackers used AI to discover and weaponize a previously unknown software vulnerability. Researchers say AI-assisted cyberattacks could evolve into semi-autonomous operations capable of adapting and scaling much faster than traditional hacking methods. These developments are increasing pressure on enterprises to strengthen AI governance before autonomous systems become deeply integrated into critical infrastructure and business operations.
Industry leaders believe agentic AI could eventually transform enterprise productivity by automating workflows, coordinating systems, and improving operational efficiency. However, many experts argue that trust will become the defining factor in adoption. Businesses will need clear oversight systems, audit trails, human intervention controls, and strong cybersecurity protections to ensure autonomous AI behaves safely and predictably. The broader debate reflects a growing realization that the future of AI is no longer just about model capability, but about whether organizations can securely govern increasingly independent digital agents operating at machine speed.
