The rapid rise of Frontier AI models is creating a new challenge for security teams, with experts warning that vulnerability discovery and exploitation are becoming increasingly automated.
According to Commvault, advanced AI systems are capable of identifying vulnerabilities at unprecedented speed, potentially reducing the window defenders have to remediate critical weaknesses.
Recent industry research suggests AI-powered security models can uncover significantly higher numbers of vulnerabilities than traditional approaches, while AI-assisted exploitation techniques are enabling attackers to move faster once flaws become public.
The result is a growing need for organisations to focus not only on prevention, but also on resilience and recovery.
Nick Patience, Vice President and AI Practice Lead at Futurum Group, said the security industry must adapt to the changing threat environment.
“While a rigorous patching strategy remains critical, the key now is also making sure readiness, resilience, and clean recoveries are top priorities,” he said.
Commvault has recommended four key actions for organisations:
- Assess whether recovery environments can withstand AI-driven attack cycles.
- Establish immutable and isolated recovery copies through air-gapped architectures.
- Prioritise recovery of systems critical to maintaining a minimum viable company.
- Automate recovery validation and conduct regular testing in cleanroom environments.
The guidance reflects a broader shift occurring across the security industry as organisations acknowledge that preventing every breach may become increasingly difficult in an AI-accelerated threat landscape.
Of particular concern is the growing reliance on AI-enabled business services. Security leaders are being encouraged to include model repositories, vector databases and agentic workflows in business continuity planning alongside traditional applications and infrastructure.
Commvault argues that operational resilience frameworks such as Resilience Operations (ResOps) will become increasingly important as organisations seek to validate recovery readiness before an incident occurs rather than during one.
