As technology in commercial aviation advances, experts see a double-edged sword: With more airline operations being handled in the digital space, the likelihood of damaging cyberattacks also increases.
And with artificial intelligence (AI) set to revolutionize the industry, cybersecurity professionals are urging caution on its rollout.
Although AI applications could improve aircraft performance, flight scheduling, and weather prediction models, they also come with potentially devastating setbacks.
“AI plays a pivotal role in both creating and mitigating cyber security vulnerabilities,” security expert Chris Cooley told The Epoch Times by text.
“On the offensive side, attackers use AI to automate tasks such as scanning for vulnerabilities or launching sophisticated phishing campaigns. For example, AI can help develop malware that evolves to avoid detection, posing a significant threat to aviation systems.”
However, Cooley also said AI is a “game changer” in terms of defense.
“It can help monitor vast amounts of data, detect real-time anomalies, and, with practical tuning, respond to threats faster than traditional systems,” he said.
“In my experience working with threat intelligence, I’ve seen how AI can enhance defensive strategies and attackers’ capabilities.”
As a certified associate chief information security officer, Cooley emphasized the importance of mitigating cybersecurity weaknesses in the airlines, which he called an “evolving” threat that could cause severe disruptions.
“Aviation relies on a vast digital infrastructure, including air traffic management systems, communication networks, and passenger services. Attacks on any of these could result in delays, grounded flights, or safety risks,” Cooley said.
For example, he said, ransomware could disrupt the booking systems for air traffic control and cause widespread operational chaos.
Bill Mann, a privacy expert at Cyber Insider who has worked with avionics, agreed and told The Epoch Times by text that further disruption to airlines from cyberattacks is “absolutely possible.”
He said that if aviation networks were breached, malign actors would most likely hold information hostage through ransomware in their customer-facing systems. This includes passenger lists and flight schedules. Mann said this could “cause chaos and ground flights.”
Mann also said it’s possible for cyberattacks to do physical damage.
“Theoretically, a very malicious person could take control of air traffic controller systems, totally disrupt the order of landings and takeoffs at an airport, and potentially cause crashes,” he said.
This is where defensive AI has a chance to shine, according to Mann. Cybersecurity systems using AI have already demonstrated the ability to learn the methodology of hackers, assist security teams with penetration testing, and create stronger networks. But Mann said the other side of this intelligence tool is “black hat AI,” which is developed specifically to exploit weaknesses in other AI systems.
“Cyber security will be decided more and more by AI in coming years,” Mann said.
AI in the Sky? Not Yet
Recognizing the challenges that cybersecurity and AI integration pose to the industry, the Federal Aviation Administration (FAA) released a Roadmap for Artificial Intelligence Safety Assurance. The report outlined the need for a structured approach to AI usage and a six-pronged method to “guide the development of methods for the safety assurance of AI in aircraft and aircraft operations.”For the moment, the FAA does not allow the use of AI in the sky.
“The FAA has not approved AI for use in existing commercial flight operations,” an agency spokesman told The Epoch Times by email.
When queried about plans to reduce cybersecurity threats in aviation, the spokesman responded: “The FAA has a comprehensive approach to protect the National Airspace System from cybersecurity threats. The agency works closely with intelligence and security experts throughout the federal government to identify and mitigate potential risks to our systems, as well as those of our partners in the private sector.”
“The large and growing digital infrastructure which supports the commercial aviation sector provides attackers with a broad and extensive cyber-attack surface,” the report states.
It also observed that a growing reliance on managed service and cloud service providers heightens the risk of indirect data breaches.
Cybersecurity threats can manifest through the exploitation of software or firmware to breach corporate networks directly or breach those belonging to service providers. This could have a downstream effect on airline operations, according to the analysis.
AI has vast potential within the industry, as it could help customers book flights and even perform aircraft maintenance, among other uses, he said.
Carvalho added that American Airlines has pilot programs to incorporate AI into nearly all elements of the company’s operations.
A New Frontier
It’s unlikely that airline passengers will see the addition of AI on the flight deck anytime soon, but that hasn’t stopped pilots from weighing in on the topic.“AI relies on data. If it doesn’t have access to data, it won’t be able to function anywhere close to the performance humans provide. Giving one system access to all this data comes with security risks,” Dingemans said.
He also pointed out that ethical considerations, regulatory challenges, and a lack of human judgment are all significant hurdles to bringing AI into the air side of operations.
“AI can significantly improve safety, efficiency, and decision-making in aviation,” Cooley said. “It can enhance predictive maintenance, optimize flight routes, and assist pilots with complex decision-making tasks. However, it’s crucial to remember that while AI can process vast amounts of data quickly, critical decisions—especially those affecting safety—must remain under human control.”
Some experts have said that machine learning, a subset of AI, is a weak spot that cyber attackers can exploit.
“Right now, it is hard to verify that the well of machine learning is free from malicious interference,” Lohn said. “In fact, there are good reasons to be worried. Attackers can poison the well’s three main resources—machine learning tools, pretrained machine learning models, and datasets for training—in ways that are extremely difficult to detect.”
This could have potential implications for machine learning’s use in aviation, especially if the programs are developed using shared data resources.
At an operational level, the report noted that Delta Air Lines uses machine learning to “price tickets, analyze passenger behavior, and provide self-service experiences.”
“As aviation becomes increasingly dependent on digital systems, combining robust cybersecurity frameworks with human oversight and a cautious approach to new technologies is crucial,” Cooley said.
