“Observability alone isn’t enough,” said Manish Ranjan, research director for software & cloud at IDC EMEA. “In today’s complex IT environments — especially with distributed AI workloads — security must be embedded as a strategic pillar, supported by governance, not treated as an afterthought.”
AI is indeed compounding existing security gaps. Companies are seeing a surge in AI-powered ransomware—up from 41% in 2024 to 58% this year—and 47% have already encountered attacks specifically targeting large language models (LLMs). Mark Walmsley, CISO at Freshfields, warned that “AI security can’t be an afterthought,” urging enterprises to adopt deep observability and rethink public cloud strategies to stay ahead of AI-driven threats, as per the statement, as per the press statement.
The AI security imperative
The architecture of the hybrid cloud itself is contributing to security lapses, according to experts. As workloads shift between on-prem and public cloud, inconsistent policies and fragmented tools create exposure. “Hybrid complexity makes fragmented controls a liability,” said Hetal Presswala, CSO at an EPC firm. “A unified security approach is critical, as varying protocols and silos increase the risk of misconfigurations and data leaks.”