5,6 A new era of cybersecurity: Why AI is both a risk and a defense mechanism

Long story short
AI is shaping emerging cyber threats and the measures that can protect against them. Standing still is not an option – safeguarding mission-critical infrastructure requires a combination of technological adoption and organizational change. It isn’t easy, but it’s necessary.
Why it matters
The rise of AI-driven cyber threats is the single largest shift in the cybersecurity space in recent memory. How businesses respond will define their success (or failure) in this area.
Emerging cyber threats in the AI age
For every new innovation in AI, a new attack surface opens up. Research from Hitachi found that data breaches from an AI-enabled attack are the number one security concern for IT leaders. Alongside sophisticated threats, the AI boom has driven an explosion of data across different geographies, jurisdictions, and third-party platforms. In this landscape, many businesses have realised the importance of data sovereignty to effectively monitor, secure and respond to risks, whilst ensuring compliance with local regulations.
As well concerns around data sovereignty, organizations are tackling the growing risks introduced by automated systems and OT environments such as telemetry. AI bots can scrape streams for credentials, API keys, and more. They can then be leveraged to enable lateral movements within compromised networks. Prompt injecting and data poisoning can likewise compromise the efficacy of LLM-based solutions.
Meanwhile, a recent study revealed that in 2024, AI-driven phishing attacks surged by 60% worldwide. Generative tools are enabling the rapid creation of deepfakes and synthetic identities, facilitating faster and more fluent attacks than ever before. Phishing is taking place at scale and is increasingly personalized, mimicking human tone and behavior. Traditional filters and static controls are ill-equipped to respond to these new challenges.
For data center leaders, this means tackling sophisticated threats across siloed, dispersed environments. They must be prepared for everything from synthetic identities used to infiltrate supply chains to automated vulnerability discovery. And they must also contend with a widening pool of cybercriminals; the UK’s National Cyber Security Centre predicts that AI tools’ availability “as a service” will embolden novice hackers by 2027.
AI-driven solutions
To keep pace with the volume and velocity of these attacks, hyperscalers, colocators, and enterprises will all need to equip themselves with AI-driven solutions  to match, making significant investments into upgrading their existing cybersecurity methodology.
To stay one step ahead of today’s cyber threats, data center leaders must prioritize proactive risk mitigation. Tools such as AI-powered anomaly detection are essential to help operators identify and neutralize threats before they escalate into full-scale attacks. Extending this intelligence down to the level of microservices enables the segmentation of architecture, helping to isolate issues and limit their blast radius. Ultimately, it may be possible to achieve systems that predict, adapt, and recover without any human intervention.
Promisingly, leaders are already confident in these solutions. Ninety-five percent of cybersecurity professionals believe that AI can increase the speed and efficiency of their response to cyber threats – from prevention and detection to response and recovery. If implemented effectively, this both improves uptime and frees up human cybersecurity specialists to focus on high-level strategy.
“Crucially, data center operators need a unified approach to cybersecurity that spans both IT and OT.”
IT/OT convergence – the missing piece of the puzzle
Crucially, data center operators need a unified approach to cybersecurity that spans both IT and OT. Through IT-OT convergence, data center operators can break down silos and achieve a single view across even the most complex environments, identifying any vulnerabilities or threats before they escalate. The alternative is undetected threats and half-baked responses – a reactive approach that’s always on the back foot.

This is especially important for hyperscalers and large enterprises that operate large, distributed infrastructures with thousands of endpoints, applications, and users. Operational leads gain better visibility of their assets, ensuring that threat actors don’t slip between the cracks of complex, distributed systems. AI systems work best when they are able to tap into data from across an organization, delivering total operational intelligence and advanced automation capabilities.
“AI systems work best when they are able to tap into data from across an organization, delivering total operational intelligence and advanced automation capabilities.”
However, it’s important to note that there can be reticence when discussing IT/OT convergence within organizations. Decision-makers who are responsible for critical assets – whether that's a BESS or a control system – might worry about the security implications of integration. Connecting systems can open up new vulnerabilities, especially if they still use legacy infrastructure that wasn't designed to operate in the age of AI-driven cyber threats.
For this reason, convergence is best undertaken in conjunction with approaches that can help tackle emerging threats, as evidenced by real-world implementations of enhanced security and ongoing support in government cloud environments. One such approach is deploying Zero Trust architecture that continuously validates identity, behavior, and intent. This removes the assumption of safety across both digital and physical infrastructure, whether you’re dealing with users, devices, or systems. It’s a robust first line of defense. Knowing that architecture has new safeguards in place will help put system owners’ minds at ease.

Case study La Molisana
La Molisana, an Italian pasta manufacturer exporting to more than 100 countries, demonstrates this multi-layered security approach. When modernizing infrastructure to support international expansion, it implemented Hitachi VSP One Block with built-in immutable snapshot functionality. Even if ransomware penetrates the network, La Molisana’s critical ERP and supply chain data remains protected, and the immutable snapshots create tamper-proof copies that attackers can’t encrypt or delete.

Find out more

Own the conversation
Ask the big question:
Will we reach the point where autonomous cybersecurity systems can address threats without the need for human intervention? Or will threats always evolve faster than they can adapt?
Disrupt your feed:
In the face of new AI threats, we need to elevate the role of information security and cybersecurity within organizations. Some companies have a CIO, but fewer have a CISO or Data Protection Officer. There is even a downward trend: just 27% of firms had a board member responsible for cyber security in 2025, compared to 38% in 2024.
Drop this fact:
78% of CISOs report that AI-powered threats are having a “significant impact” on their organizations.

12 January 2026

A new era of cybersecurity: Why AI is both a risk and a defense mechanism
Roger Lvin CEO and Executive Chairman

Addressing the data center talent crisis

How strategic partnerships will transform the data center

Think beyond the data center

Enabling AI-driven data center operations with IT/OT Convergence

Driving efficient resource use and reusing operational waste to support sustainability ambitions

Redefining cybersecurity and trust in the new data center landscape

Digital twins will rewrite the design rules for the data center

AIOps will unlock the next generation of hybrid infrastructure

12 January 2026 A new era of cybersecurity: Why AI is both a risk and a defense mechanism Roger Lvin CEO and Executive Chairman