The most exposed systems are generally those that directly depend on Global Navigation Satellite System (GNSS) – derived positioning, navigation, or timing, or use GNSS as an input to another system.
That includes navigation systems, performance-based navigation, surveillance functions and some safety and automation functions, Markus Levin, co-founder of XYO – a decentralised physical infrastructure network told Oceanblumedia.
With modern conflicts consuming nearly every region in the world, aircrafts and ships are being targeted – specifically the vessel’s navigation systems. Levin said the underlying vulnerability is the same.
Both ships and aircraft depend on GNSS for positioning, navigation, and timing, and civilian GNSS signals are relatively weak and were not designed around the assumption that they would be deliberately manipulated.
The difference is in how the consequences show up. At sea, a spoofed position can make a vessel appear to be somewhere it is not, affecting navigation, collision avoidance, port operations and other systems that depend on location. In aviation, the same problem can spread into navigation, surveillance, and safety-related functions. The aircraft may still be flying normally, but the systems supporting the crew can begin working from an incorrect position or timing reference.
Aviation is particularly sensitive to this because GNSS is embedded across multiple parts of modern flight operations, including performance-based navigation and some surveillance and safety functions.
Also Read: Hyundai Motor Group Accelerates AI Transformation
Aviation Spoofing
The seriousness of GPS/GNSS interference is becoming more serious for commercial aviation. Levin said that five years ago, most pilots and many people across the industry were much less familiar with spoofing, and regulators hadn’t started tracking it the way they do today.
European Union Aviation Safety Agency (EASA) updated its GNSS interference safety bulletin in July 2024 following analysis of recent jamming and spoofing occurrences, and it specifically describes the incidents as increasing in severity and sophistication.
The geography has changed too, since interference isn’t tied to one isolated incident anymore. There has been persistent interference affecting civil aviation, including incidents around the Korean Peninsula.
Aviation has layers of procedures and redundancy for exactly this reason, and commercial aviation hasn’t suddenly become unsafe. The concern is that GNSS interference is becoming persistent enough that airlines, pilots, air navigation providers, and manufacturers have to treat it as a normal operating condition in some regions.
That changes the risk calculation. An occasional loss of signal is something you can treat as an anomaly, but a persistent one means you have to design the system around the assumption that GNSS can sometimes be wrong.
Advancing technology and hardware is not helping.
Levin pointed out that better hardware does not change the fundamental trust model of the signal.
GNSS was designed when availability and global coverage were the primary objectives. Civilian signals were not built around strong authentication that allows a receiver to conclusively determine that every signal it is receiving is genuine.
Modern aircraft have much better sensors, processors and navigation computers than earlier generations, but those systems still need inputs. If one of those inputs is false and the system has insufficient independent evidence to reject it, more processing power does not solve the problem.
We have built highly sophisticated systems on top of a location signal that was never designed to provide cryptographic proof of its own authenticity. The industry needs to move from asking only ‘Where does the system say we are?’ to also asking ‘What evidence do we have that this position is genuine?’
Sophisticated Attacks
navigation failure.
With jamming, the signal simply disappears or becomes unusable. Spoofing is different, because the aircraft keeps receiving what looks like usable positioning information, and that information happens to be wrong.
The result can be inconsistencies between different navigation sources, unexpected position changes, degraded navigation performance, or alerts from systems that depend on reliable positioning.
The aircraft doesn’t need to behave dramatically differently for the event to matter, because a false input gets folded into other calculations and creates a chain of secondary effects.
Creating a false signal environment is not especially hard, because civilian GNSS was not designed around strong authentication. Making a false position remain credible across an entire aircraft’s independent navigation sources is a separate and harder problem.
A modern aircraft does not depend on one piece of information. It can combine GNSS with inertial reference systems and other navigation inputs. When those sources disagree, the inconsistency can become detectable.
The more independent evidence an aircraft has about its position, velocity, and movement, the harder it becomes for a single manipulated input to establish itself as ground truth.
Also Read: AI Agents Could Outnumber Human Workers Globally by 2029
Exposed Systems
Levin believes the most exposed systems are generally those that directly depend on GNSS-derived positioning, navigation, or timing, or use GNSS as an input to another system.
That includes navigation systems, performance-based navigation, surveillance functions and some safety and automation functions.
A useful example comes from the Incheon Flight Information Region. Between late May and early June 2024, roughly 500 flights and hundreds of vessels in the region reported GPS problems, and some of those events were also linked to false ground-proximity warning system alerts.
Mitigation measures prevented serious safety consequences, but the scale of the reports showed how broadly an interference event can affect aviation operations. South Korea has since reported to ICAO that interference in the Incheon FIR affected more than 4,400 flights from over 20 countries between October 2024 and February 2025.
There have also been repeated reports of spoofing affecting aircraft operating around the Middle East, and more recently in Europe.
Other Navigation Systems
They are important because resilience comes from diversity, Levin said.
Inertial navigation is one of the most obvious complements because it does not depend on receiving an external satellite signal. Its limitation is that errors accumulate over time, which means it benefits from being periodically corrected by another independent source.
That is where terrestrial signals, visual navigation, radar, radio-based navigation and other sensors can become useful. Computer vision can provide another source of environmental information, while terrestrial systems can provide independent timing or positioning references. The goal is to prevent GNSS from being the only trusted source.
The expert recommended that the first step is to understand that GNSS can sometimes be wrong. That sounds simple, but it is an important change in how systems are designed.
Aircraft should continue using GNSS because it is extremely valuable, but critical systems should increasingly cross-check it against independent sensors and navigation sources. Operators should have clear procedures for degraded or compromised GNSS environments, and crews should be trained to recognize the difference between a signal outage and potentially misleading positioning data.
The second step is technological. We need more resilient navigation architectures that combine GNSS, inertial systems and independent sources rather than treating GNSS as an unquestioned ground truth.
The longer-term fix is treating a coordinate as a claim rather than a fact. Right now, a system receives a position and accepts it, when it should be able to check what backs that position instead, cross-referencing other sensors, other independent sources, and whether it lines up with where the aircraft already was.
Some form of independent verification is what makes that checking possible without trusting any single input by default.