Rethinking Risk Management in the Era of AI

A recent survey by Gigamon reveals a startling increase in cyberattacks, with a reported 17% rise in breaches over the past year, as nearly 55% of organizations faced direct attacks. This growing threat landscape, driven by advancements in artificial intelligence, presents significant challenges for security teams that are grappling with visibility issues. The findings underscore the urgency for businesses to evaluate their cybersecurity strategies, particularly concerning public cloud vulnerabilities.

The survey indicates that AI is accelerating the exposure of critical security weaknesses. With global investments in AI projected to exceed $200 billion by 2025 and reach approximately $750 billion by 2028, industry leaders acknowledge the pressing need to redefine risk management in this evolving context. A concerning 91% of security leaders admitted to making compromises that weaken their hybrid cloud security, primarily due to gaps in visibility.

Significantly, the report highlights a surge in AI-driven ransomware incidents, with 58% of respondents reporting an increase in attacks compared to 41% in the previous year. Nearly half of the surveyed organizations reported threats directly targeting their large language model deployments. In addition, adversarial AI tactics are flooding operational environments with sophisticated phishing attacks, deepfake impersonations, and obfuscation techniques that conceal data exfiltration.

The escalating volume of network data poses further challenges, with one in three organizations noting that their data volumes have more than doubled in just two years, largely due to AI workloads. Traditional security tools are increasingly ineffective under this pressure, prompting 46% of security and IT leaders to prioritize managing AI-related threats as a critical focus for new technology deployments.

Rising Concerns Over Public Cloud Security

Once viewed as a symbol of efficiency and scalability, the public cloud is now regarded as highly risky by 70% of survey respondents. More than half of these organizations are refraining from AI deployments within public cloud environments, citing concerns about intellectual property security and governance risks. Additionally, a significant portion of respondents is contemplating reverting data and workloads back to private cloud or on-premises solutions.

This shift indicates a greater awareness that speed is irrelevant without adequate control. As the attack surface broadens with AI technology, security teams are re-evaluating not only the environments in which their data resides but also the levels of visibility and manageability they maintain.

Deep Observability as a Strategic Response

A key insight from the survey is that nearly half of organizations—47%—lack comprehensive visibility into their hybrid infrastructure, including lateral traffic and AI-generated data. Real-time insight into all data in motion has emerged as a top security priority for 64% of respondents, highlighting a critical gap in current security measures.

Deep observability is proposed as a viable solution, integrating traditional performance metrics with advanced network telemetry like packet data and flows. This approach offers unparalleled clarity for monitoring complex environments. Over 88% of leaders regard deep observability as essential for securing AI deployments, with discussions of its importance reaching board-level considerations.

Challenges Faced by CISOs

Chief Information Security Officers (CISOs) find themselves central to this dynamic but often lack the necessary authority and resources to implement effective cybersecurity measures. A staggering 97% of CISOs recognize having to make security trade-offs, frequently due to fragmented tools and unmonitored AI deployments. However, the overwhelming majority—86%—assert that access to granular packet-level data is crucial for improving threat detection and response capabilities.

The stakes are high, as 80% of CISOs now equate cybersecurity risks with financial and legal liabilities, raising the cost of inaction to unprecedented levels. This evolving landscape necessitates a fundamental shift in how security is approached, emphasizing proactive measures over reactive responses.

Conclusion and Future Outlook

The hybrid cloud is not disappearing, nor is the relevance of AI diminishing. As these technologies continue to develop, businesses must recalibrate their security frameworks to transition from reactive to proactive management. By enhancing visibility and comprehension of their environments, organizations can navigate risks more effectively. The concept of deep observability is not merely a technical upgrade but a critical strategic need for organizations aiming to thrive amidst future challenges. For more insights, security leaders can explore the findings from the Gigamon 2025 Hybrid Cloud Security Survey and accompanying reports focused on adapting security strategies in the age of AI.