In the last few weeks, we’ve seen waves of information technology infrastructure outages and failures worldwide, especially across Western Europe and the United States. All of us at McGill likely experienced last week’s Microsoft outage to some degree, which affected public and private institutions worldwide and lasted over 36 hours. Other examples include the grocery fridge outages across European countries, including Sweden, Belgium, Estonia, and France; the recent Binance hack, which led to a $6.36 billion loss for the international cryptocurrency broker; the American Federal Aviation Administration (FAA) outage last week, which caused an extended grounding of nearly all outgoing flights across the USA; and the outage of the American National Security Agency’s (NSA) secure internal communications networks this week.
Given current geopolitical instability, particularly the US-Iran war, many observers are quick to ask whether these attacks are part of a coordinated attempt to destabilize the US and its allies at home. Others point to the increasing capabilities of AI models in coding, and a history of hacks using AI undertaken by bad actors, as the culprit; the proliferation of high-powered AI systems, they argue, is responsible for empowering teams of bad actors at any scale to wage cyber warfare against information technology systems with ease. This might be true for some outages: the French claim the Iranian Revolutionary Guard Corps is responsible for its refrigeration system outages (though Iran rejects this claim), and investigators claim North Korea used AI as an integral tool for hacking Binance’s crypto reserves. The US Department of Homeland Security also claims that Iran has been conducting cyber warfare on its infrastructure, such as water systems. However, it’s clearly not been true for all, with leaders at the FAA, the NSA, and Microsoft claiming that outside forces had nothing to do with their outages and, in the case of the first, that system upgrades led to infrastructure failures.
This whole debacle has led to arguments over whether increased paranoia about the safety of core infrastructure systems in advanced nations is warranted, or whether these co-occurrences are mere coincidence. American President Donald Trump weighed in, leaning towards the latter interpretation: “It’s life,” he said. Others online have formulated intricate conspiracies, blaming Iran or Israel.
However, as Liberal Arts students, we must think across the different spheres that influence and shape the context in which these infrastructural failures occur; we have to think in systems. The relevant systems include ones we’ve touched on: domestic politics, geopolitics, and technological innovation. The co-occurrence of instabilities in these three societal superstructures reshapes the environment in which our pre-existing infrastructural systems exist. Plainly, the foundation underlying IT infrastructure—government capacity, availability of capable AI, geopolitical instability—is changing in ways that make all sorts of vulnerabilities more likely to be exploited or to fail. Thus, rather than mere coincidences, these infrastructural failures are better understood as structural incidents—the heightened rate of clustered outages in digital systems is the new norm in an era of converging instabilities.
AI poses a direct threat to practically all of our digital infrastructures. As one backend systems engineer told me:
“I think the negative truth in all of it is that everything is under a baseline cyber attack […], but that’s been true for years. It’s only gotten worse this year, with AI making it much easier to build tools.”
This was very much the case in the Binance hack, where sophisticated North Korean hackers leveraged AI tools to execute the most expensive theft in history.
However, advances in AI also flow to those who attempt to secure and maintain systems. William Leduc, a seasoned IT systems engineer, offered me the following metaphor for the multi-faceted nature of the situation:
“The usage of AI in security is basically the introduction of easier to use inspection tools in software, but on the flip side, is also like giving X-ray goggles to crackheads looking for copper wires.”
Focusing on the impact of AI on security, moving fast to implement it in security systems might itself pose an indirect safety risk. As many IT security systems prioritize long-term stability over dynamism, they might be ill-suited to the fast-paced hot-patching of new code or layers of protection that defending against a much more rapidly evolving cybersecurity environment requires. We see a similar dynamic in the case of the FAA, where a far-outdated system running on decades-old routers was taken down by installing newer routers that operated at such high speeds relative to the old ones that they overwhelmed the network. Thus, the old paradox of security systems engineering becomes all the more relevant: both increased attempts to attack and to secure a network system may be detrimental to its stability.
Even as agencies scramble to modernize their hardware, Washington has dismantled the primary body tasked with coordinating that defence. The Trump administration has gutted the Cybersecurity and Infrastructure Security Agency (CISA). This government body was previously responsible for coordinating the American tech industry’s response to critical bugs, vulnerabilities, and exploits in core infrastructure software and hardware. While nations around the world that rely on American technologies in their networks are now more vulnerable than ever, the mechanisms they could previously rely on to defend themselves have had their capabilities severely hampered.
CISA’s weakness feeds directly back into the earlier issues. Within the context of AI quickly finding code vulnerabilities and exploits, the absence of a central clearinghouse to coordinate structural changes and strategies leaves private and public institutions relying on precisely the kind of patchwork of hot-fixes that are increasingly themselves the cause of systemic instability.
Our direct experience with the Microsoft outages is a harbinger of what’s to come in our lifetimes if the superstructural scaffolding that underpins our digital infrastructure continues to be stripped away. Unless Canada and other nations reliant on American tech build real digital sovereignty—and rebuild the institutional guardrails Washington has abandoned—the resounding emptiness of “it’s life” will become our everyday experience.