Agentic AI in Operational Technology: Risks and Governance
In a recent discussion, Gartner’s Wam Voster highlighted the critical cybersecurity risks associated with integrating agentic AI into operational technology (OT) environments. As factories increasingly depend on AI to manage production processes, the potential for harmful decision-making rises. Voster warned that automated AI systems adjusting key variables such as temperature could lead to severe consequences, including catastrophic equipment failures and threats to human safety. He noted that a misconfiguration in AI decision-making could escalate a temperature increase from 2 degrees to a dangerous 200 degrees.
Moreover, Voster emphasized the importance of robust governance and risk management strategies to mitigate these risks, advocating for better asset visibility and addressing human error, which frequently triggers cyber-physical security incidents. These insights underscore the need for careful oversight in deploying AI within critical infrastructure, ensuring safety and resilience across industrial operations.
Keywords: Agentic AI, Governance, Risk Management, Operational Technology, Cybersecurity, AI Systems, Industrial Safety, Critical Infrastructure.
