Managing Risks and Optimizing the Value of AI, GenAI & Agentic AI

51% of organizations say human oversight is needed in AI governance due to the speed at which attackers can adapt.

62% of organizations say it is difficult to minimize model and bias risks in language model development.

56% of organizations report challenges in managing AI user risks, including the unintended spread of misinformation.

52% of enterprises have fully or partially deployed GenAI.

47% of organizations say their AI models can learn robust norms and make safe decisions autonomously.

40% of organizations report errors in data inputs ingested by AI.

59% of organizations say AI makes it more difficult to comply with privacy and security regulations.

Only 1 in 5 enterprises reach AI maturity, where AI in cybersecurity activities is fully deployed and security risks are assessed.

51% of organizations say AI is effective in reducing the time to detect anomalies or emerging threats.

Only 41% of organizations have AI-specific data privacy policies in place.

58% of organizations say prompt or input risks (e.g., misleading, inaccurate, or harmful responses) are very or extremely difficult to minimize.

48% of organizations rate AI as effective in threat detection and hunting for deeper insights and reducing manual workload.

45% of organizations cite errors in AI decision rules as a top barrier to operational effectiveness.

43% of enterprises have adopted a risk-based strategy to govern AI systems.

79% of organizations have not yet reached full AI maturity in cybersecurity.