As organisations race to adopt artificial intelligence, the world’s cybersecurity leaders are being forced to rethink their entire defensive playbook. New research from Gigamon reveals that as AI reshapes how businesses innovate and how adversaries operate, chief information security officers (CISOs) are recalibrating their risk strategies to meet a fast-evolving threat landscape.

The global study, “CISO Insights: Recalibrating Risk in the Age of AI,” surveyed more than 200 CISOs across Australia, France, Germany, Singapore, the UK and the US. It found that 86 percent regard metadata as essential to achieving the complete visibility required to protect today’s hybrid cloud environments. As network traffic grows exponentially, metadata offers a scalable way to surface critical signals from within complex data flows, helping security teams identify and respond to threats faster.

Visibility at the Centre of Cyber Strategy

Visibility has emerged as the top priority for CISOs, with real-time threat monitoring now viewed as the linchpin of effective cyber defence. Yet nearly all respondents—97 percent—admitted to making compromises in visibility, tool integration or data quality. These gaps leave organisations vulnerable to the blind spots that attackers exploit.

Chaim Mazal, chief AI and security officer at Gigamon, said the rise of AI-driven ransomware, social engineering and shadow AI is reshaping the way leaders approach defence. “AI remains both one of the biggest challenges and most exciting opportunities for CISOs,” he said. “While AI introduces new attack vectors, it also provides powerful tools to restore visibility and control, strengthening how organisations structure and resource their defences.”

Cloud Complexity and AI Data Volumes

The report highlights how data volumes driven by AI applications have nearly doubled, expanding the attack surface across virtual, cloud and containerised environments. Seventy-five percent of CISOs said the public cloud now poses a greater security risk than any other environment, prompting many to rethink data storage strategies. Nearly three-quarters are considering repatriating workloads to private cloud, while more than half are reluctant to use AI in the public cloud due to intellectual property and compliance concerns.

The issue has become a board-level priority, with 70 percent of respondents saying public cloud security is now discussed regularly at executive level.

Securing AI Through Better Data Insight

One in five CISOs said they are not confident their existing tools can handle the surge in network data created by AI systems. Traditional log-based solutions are struggling to detect and respond to AI-enabled attacks, pushing leaders to prioritise visibility and data enrichment. Over the next 12 months, more than half plan to leverage network and application metadata to enhance the effectiveness of existing tools, while 46 percent are working to ensure complete visibility across all data in motion. Around one-third are implementing guardrails around large language models to mitigate exposure to emerging risks.

AI’s Dual Role: Threat and Ally

As the number of data breaches climbs—up 17 percent year-on-year—CISOs are feeling the pressure. Nearly half said they personally shoulder accountability when breaches occur, and many cited the toll of stress and burnout on their teams. With skills shortages and shrinking budgets, AI is increasingly being seen as both a necessity and an opportunity. Forty-five percent of CISOs are already using AI tools to enhance their teams’ capabilities, while 73 percent see AI as a way to offset declining headcount.

AI is also helping to upskill junior analysts, automate root-cause analysis and improve threat visibility, allowing teams to respond with greater speed and precision.

Deep Observability as the Foundation for Secure AI

The research found that 82 percent of CISOs now consider deep observability—bringing together network-derived telemetry and log data across hybrid cloud environments—a foundational requirement for secure and efficient AI deployments. By unifying visibility across all layers of infrastructure, deep observability enables organisations to detect anomalies earlier, ensure AI models are trained on trusted data, and strengthen overall cyber resilience.

The findings make one thing clear: as AI becomes integral to business operations, visibility and data integrity will define the next era of cybersecurity leadership.