As artificial intelligence systems become more prevalent in enterprise environments, organizations are grappling with new categories of security and operational risks that traditional frameworks struggle to address. According to industry data, 79% of organizations now have some form of AI agent in production, with 96% planning expansion.

The rapid deployment has created significant demand for cybersecurity expertise. Security engineers are seeing increased demand as artificial intelligence generates large volumes of new code and creates novel security concerns that require specialized knowledge to address.

Government agencies are responding to these emerging risks. The UK's AI Security Institute, staffed by alumni from major AI companies including OpenAI and Google, is developing approaches to identify and mitigate AI-related dangers. The institute is serving as a model for other countries working to understand and manage AI risks.

A particularly challenging issue involves autonomous AI agents that can modify infrastructure configurations, restart services, or reroute network traffic in response to detected problems. Industry experts report that these agents can inadvertently trigger cascading system failures when they take actions without full awareness of broader system conditions.

Research from the AI Incidents Database shows reported AI-related incidents rose 21% from 2024 to 2025, though experts believe actual numbers may be higher since many organizations lack incident classification systems that capture autonomous agent actions as root causes of system failures.

Gartner predicts that 33% of enterprise software will include agentic AI by 2028, but separately warns that 40% of those projects may be canceled due to inadequate risk controls. The prediction highlights the tension between rapid AI adoption and the development of appropriate governance frameworks for managing these new technologies in production environments.