AI could expose cloud weaknesses

YARMOUTH, Maine — Jan. 28th marks the 20th annual Data Protection Day, and as more complex threats arise from the development of artificial intelligence (AI), companies must look to developing their infrastructure to keep up, say industry experts. 

The time for relying on large service providers like AWS and Google will be a thing of the past as companies move to better secure their data, says Roger Brulotte, CEO, Leaseweb Canada. 

“In 2026, the organizations that lead on data privacy will be the ones that diversify deliberately, build for control instead of convenience, and stop assuming that the biggest provider is automatically the safest choice,” he said. “2026 is the year infrastructure strategy becomes inseparable from data protection and privacy strategy." 

Fortunately, today’s organizations have far more tools to protect sensitive information than they did two decades ago. From multi-cloud deployments and on-premises data centers to edge computing environments, companies now have a wider range of options to improve security, performance and control. 

Richard Copeland, CEO of Leaseweb USA, says the shift is rooted in a move toward “verifiable isolation rather than trust in the provider’s perimeter,” though even that approach has its challenges. 

“AI’s shift from simple automation to agentic workflows is exposing the weaknesses of large, multi-tenant hyperscale environments,” he said. “As AI continues to become more and more critical, the less tolerant organizations are of unpredictable billing, noisy-neighbor issues, opaque GPU allocation, or cascading failures triggered by one overloaded service. Likewise, attackers are evolving too, using AI to exploit precisely those blind spots.” 

As a result, Copeland said organizations are gravitating toward regional data centers and bare-metal infrastructure, where they can gain cleaner environments, transparent performance metrics, and a reduced blast radius in the event of a breach or outage. 

Russ Ernst, CTO of Blancco, says that when it comes to AI and data collection, adopting a “less is more” approach is increasingly important. 

“The proliferation of AI also has the potential to threaten data privacy because it often relies on large datasets, which can encourage organizations to collect more personal data than necessary,” he said. “Without strict limits, sensitive information may be used for purposes customers never consented to, violating privacy expectations and regulations.” 

Genetec recently shared a list of best practices for data protection that you can find online at Security Systems News.