The importance of critical infrastructure in maintaining economic stability and mitigating supply chain disruptions is highlighted every time a part of that infrastructure fails — such as the collapse of Maryland’s Francis Scott Key Bridge. That disaster was manmade. Many such failures result from natural events like tornadoes, hurricanes, floods, and ice storms. What is most troubling today is that critical infrastructure is becoming a deliberate target of nefarious actors. The U.S. Department of Homeland Security (DHS) defines critical infrastructure as “the vast network of highways, connecting bridges and tunnels, railways, utilities and buildings necessary to maintain normalcy in daily life.” The DHS adds, “Transportation, commerce, clean water and electricity all rely on these vital systems.” For years, I have advocated the use of artificial intelligence (AI) to help protect critical infrastructure. Benedict Macon-Cooney, Chief Policy Strategist at the Tony Blair Institute for Global Change, goes a step further. He insists AI should also be considered critical infrastructure. Macon-Cooney writes, “Building government artificial intelligence capability is as important as roads and rail.”[1]
Artificial Intelligence and Critical Infrastructure
There are two primary ways artificial intelligence can benefit critical infrastructure. First, AI can make the operation of critical infrastructure more efficient. Second, it can help provide better cybersecurity for critical infrastructure.
Improved Efficiency
Recently, there has been a lot of press about the need to upgrade America’s power grid and make it more efficient. Peter Kohnstam, Sales Director of U.S. generation and transmission at Nexans, writes, “Electrical grids across the world are straining under the pressures of contemporary life. From phones to cars, the electrical grid energizes our lives, but it is far from modern. In the U.S., the last major overhaul of the system was in the middle of the 20th century and some components of the grid are approaching 100 years old. In that same time, the world has become more reliant on electricity.”[2] There has also been a lot of press about how much energy AI requires. Journalist Neal Freyman reports, “Over the past few decades, US demand for electricity has been pretty flat. But the horizontal chart went fully vertical as the excitement around AI set off a land grab for data centers, which are massive buildings housing the computing infrastructure that trains large-language models, like the one behind ChatGPT.”[3] Ironically, AI can help address both concerns. Marc Spieler, Nvidia’s head of energy, suggests “the future of grids could be akin to an autonomous car.”[4] Using powerful AI chips and the right data, power grids could “make real-time decisions … and then … update the model, share what it’s learned, and ingest what other people have learned.” AI can also be used to address its own power-hungry appetite. Every efficiency effort will be needed in the years ahead.
To help cities protect valuable water resources, planners can leverage digital twin technology to ensure the most efficient flow of water. Journalist Matt Simon observes, “The bigger a city gets, the more wasteful it gets with water, even though its energy usage becomes more efficient.”[5] That fact alone should encourage municipal planners to use AI to help make systems more efficient. Water leaks are also a major concern. Journalist Nick Pipitone reports, “Today’s advanced sensor tech tracks recorded water usage patterns over time, allowing it to detect anomalies in pressure and flow rates that indicate possible leaks. The AI systems continuously learn from vast amounts of data, enabling them to differentiate between normal pipe operating variations and leak-related abnormalities. They identify leaks in real time, improving detection efficiency.”[6] In every critical infrastructure area — power, water, roads, rail, and ports — the story is the same: AI can help make systems more efficient and cost effective.
Artificial intelligence is becoming so important that Macon-Cooney predicts AI itself will eventually be considered critical infrastructure. He explains, “A comprehensive digital infrastructure — which includes national computing power, a distributed cloud, and an interoperable set of applications and machine-readable legislation — will be as important to a country as roads, rail, and public water supply. … More and more countries will accelerate the building of such nationwide digital architectures, allowing them to deliver more AI-powered responsive services that cater to the individual and help the population at large. … Bold governments will be making this move — and they will be examples to follow for the rest of the world.”
Better Cybersecurity
As noted above, cybersecurity is big concern when it comes to critical infrastructure. The Digital Watch staff notes, “In today’s increasingly interconnected world, protecting critical infrastructure systems is paramount. AI can be a valuable ally in this endeavor. But its misuse can present a risk to critical infrastructure systems.”[7] The staff goes on to identify several ways AI can help safeguard critical infrastructure. They note, “AI can strengthen the security of critical infrastructure by employing advanced threat detection and response mechanisms. AI-powered cybersecurity systems constantly scan network activity, system records, and user behavior for possible threats and abnormalities. These systems use machine learning algorithms to analyze enormous volumes of data, identify patterns linked to well-known and newly discovered cyberattacks, and take prompt, proactive measures to reduce risks. Additionally, AI-based intrusion detection and prevention systems have the capacity to recognize and prevent harmful activity and intercept attempts at unauthorized access. These systems continuously improve their capacity to identify sophisticated and developing threats by learning from previous data, adapting to new attack strategies, and collecting new information.”
Cyberattacks on critical infrastructure occur more often than most people realize. The Port of Los Angeles alone reports it stopped 750 million cyber-intrusion attempts in 2023.[8] Some of these attacks are made my criminal groups looking to lock up systems and so they can demand ransoms to release them. Even more troubling are state-sponsored hackers trying to control or disable the critical infrastructure of other countries. The Digital Watch staff explains, “It is important to acknowledge that malicious actors can … exploit AI. They can utilize AI to create sophisticated evasion techniques that enable them to get around conventional security measures, for instance, by creating undetected malware variants or mimicking legitimate user behavior. Additionally, they can use AI to automate attacks, which would use algorithms to locate targets and initiate and modify tactics as needed. Lastly, attackers can conduct data poisoning attacks by altering AI training data to introduce biases or weaknesses in the systems. To mitigate these risks, organizations need to implement robust cybersecurity measures, regularly update their systems, and deploy multi-layered defense strategies. Combining human expertise and oversight with AI capabilities promises more effective defenses.”
The staff concludes, “The strength of AI in anomaly detection is particularly valuable for critical infrastructure cybersecurity. By developing behavior models, AI systems can spot deviations and anomalies that can be signs of security breaches or cyberattacks. This approach enables early detection of previously unseen attack vectors.” Iran, North Korea, China, and Russia are all known to conduct malicious cyberattacks against other nations. Recently, a group of hackers linked to the Russian military intelligence unit known as Sandworm carried out attacks against critical water infrastructure. “[The group] claimed responsibility for directly targeting the digital systems of water utilities in the United States and Poland as well as a water mill in France, flipping switches and changing software settings in an apparent effort to sabotage those countries’ critical infrastructure.”[9]
Concluding Thoughts
A report from the RAND Corporation concludes, “AI is transformative technology and will likely be incorporated broadly across society — including in critical infrastructure. AI will likely be affected by many of the same factors as other information age technologies, such as cybersecurity, protecting intellectual property, ensuring key data protections, and protecting proprietary methods and processes.”[10] The Digital Watch staff adds, “Critical (information) infrastructure protection (CIP) is ever more important because critical infrastructures depend increasingly on networks linked to the Internet. Many vital parts of global society — including industries such as energy, water, and finance — are becoming more and more dependent on the internet and other computer networks as an information infrastructure. While allowing for resource optimization, this also leaves them at the risk of a cyberattack or an internet fallout.” This sets up AI vs AI confrontations in the area of critical infrastructure. It’s a battle the good guys can’t afford to lose.
Footnotes
[1] Benedict Macon-Cooney, “AI Is Now Essential National Infrastructure,” Wired, 26 December 2022.
[2] Peter Kohnstam, “2023: The Year to Overhaul Electricity Grids for a Better Future,” SupplyChainBrain, 4 January 2023.
[3] Neal Freyman, “The power grid has an AI problem,” Morning Brew, 24 March 2024.
[4] Hayden Field and Grace Donnelly, “Here’s how AI could transform the US power grid,” Tech Brew, 26 August 2022.
[5] Matt Simon, “The City of Tomorrow Will Run on Your Toilet Water,” Wired, 12 February 2024.
[6] Nick Pipitone, “AI Is Learning the Secret Language of Pipes To Detect Water Leaks,” Propmodo, 24 March 2024.
[7] Staff, “AI and critical infrastructure,” Digital Watch.
[8] Staff, “Port of Los Angeles Stopped Millions of Cyber-Intrusions in 2023,” SupplyChainBrain, 23 April 2024.
[9] Andy Greenberg, “Hackers Linked to Russia’s Military Claim Credit for Sabotaging US Water Utilities,” Wired, 17 April 2024.
[10] Staff, “Artificial Intelligence and Critical Infrastructure, Homeland Security News Wire, 9 April 2024.
