Microsoft chief executive Satya Nadella told reporters on Tuesday that artificial intelligence is generating both opportunities and new security challenges for the connected world, and that a large portion of critical infrastructure may be "unpatchable." He made the comments while downplaying a recent Outlook outage that affected some users.
What Nadella Said About AI and Infrastructure Security
During a press briefing, Nadella said AI-driven tools are being adopted at unprecedented speed across sectors such as energy, transport, health and finance. He warned that many legacy systems that run power grids, water treatment plants and other essential services were not built with modern security updates in mind.
"We see a growing number of attacks that leverage AI to find and exploit vulnerabilities that were previously considered unpatchable," Nadella said. "If we cannot patch these systems quickly, the impact on public safety could be severe."
Why the Outlook Outage Was Mentioned
The remarks came after Microsoft confirmed a brief outage of its Outlook email service on Monday, which lasted about 45 minutes for a subset of users in North America and Europe. The company said the incident was caused by a misconfiguration in a load-balancing component and that all services were restored within an hour.
Nadella downplayed the event, noting that "the outage was limited in scope and did not affect the core security of our platforms." He used the moment to pivot to a broader discussion about how AI is reshaping both productivity and risk.
Background: AI's Rise in Cybersecurity
Artificial intelligence has become a double-edged sword in the cyber realm. On one side, AI-based detection systems can analyse massive data streams to spot anomalies faster than human analysts. On the other, threat actors are using generative models to craft phishing emails, automate vulnerability scanning and even develop malware that can adapt to defenses.
Industry reports from the past two years show a steady increase in AI-enhanced attacks. The 2023 Global Threat Landscape report from the Cybersecurity and Infrastructure Security Agency (CISA) noted a 34 % rise in AI-generated phishing attempts compared with the previous year.
What 'Unpatchable' Means for Critical Infrastructure
Legacy infrastructure often runs on operating systems and control software that are no longer supported by vendors. When a vulnerability is discovered, manufacturers may not be able to release a patch because the hardware cannot accommodate new code, or because the system is isolated from the internet and cannot receive updates without risking downtime.
In such environments, AI can accelerate the discovery of zero-day exploits, vulnerabilities that are unknown to the vendor. Because the systems cannot be patched quickly, the window of exposure may extend for months or even years.
Impact on Different Sectors
Energy and Utilities
Power grids rely on supervisory control and data acquisition (SCADA) systems that were often installed in the 1990s. Many of these devices lack modern authentication mechanisms, making them attractive targets for AI-driven attacks that can manipulate load balancing or cause outages.
Transportation
Rail signalling and air traffic control networks are similarly entrenched. An AI-crafted exploit could interfere with train routing or aircraft communication, potentially leading to safety incidents.
Healthcare
Hospitals use a mix of legacy medical devices and newer cloud-based records systems. If AI can bypass authentication on older devices, patient data and even treatment delivery could be compromised.
Microsoft's Response and Initiatives
Microsoft has been investing heavily in AI-driven security tools under its "Microsoft Security Copilot" and "Defender for Cloud" suites. Nadella highlighted that the company is working with governments and industry groups to develop standards for secure AI deployment.
He also referenced the company's "Zero Trust" framework, which assumes that every request, whether inside or outside the network, must be verified. "Zero Trust combined with AI-enhanced monitoring gives us a better chance of detecting anomalous behaviour before it becomes an incident," he said.
What Regulators Are Doing
In the United States, the Department of Homeland Security's Cybersecurity and Infrastructure Security Agency has issued guidance urging critical-infrastructure owners to adopt AI-based threat-intelligence platforms. The European Union's Cybersecurity Act, updated in 2023, now requires member states to assess AI-related risks as part of their national cybersecurity strategies.
What Comes Next
Microsoft plans to release a set of AI-powered security updates for its Azure cloud services later this year, aimed at helping customers patch vulnerable workloads more quickly. The company also announced a partnership with several SCADA manufacturers to develop retrofit kits that can add AI-based intrusion detection to older equipment.
Industry observers say the next few months will be critical as regulators finalize AI-specific cybersecurity standards and as more organizations adopt AI tools for both productivity and defense.
Practical Takeaways for Readers
- Businesses running legacy systems should audit their security posture and consider AI-enabled monitoring solutions.
- Organizations should adopt a Zero Trust architecture to limit the blast radius of any potential breach.
- Stakeholders in critical sectors should stay informed about upcoming regulatory requirements related to AI and cybersecurity.
While the Outlook outage was a temporary inconvenience, Nadella's comments signal that the broader conversation about AI's impact on security is gaining urgency at the highest levels of the tech industry.