AIS Data Vulnerability Indicated by a Spoofing Case-Study

Abstract

This paper takes a close look at the landscape of the Automatic Identification System (AIS) as a major source of information for maritime situational awareness (MSA) and identifies its vulnerabilities and challenges for safe navigation and shipping. As an important subset of cyber threats affecting many maritime systems, the AIS is subject to problems of tampering and reliability; indeed, the messages received may be inadvertently false, jammed, or intentionally spoofed. A systematic literature review was conducted for this article, complemented by a case study of a specific spoofing event near Elba in December 2019, which confirmed that the typical maritime AIS could be easily spoofed and generate erroneous position information. This intentional spoofing has affected navigation in international waters and passage through territorial waters. The maritime industry is neither immune to cyberattacks nor fully prepared for the risks associated with the use of modern digital systems. Maintaining seaworthiness in the face of the impact of digital technologies requires a robust cybersecurity framework.

1. Introduction

Maritime industry stakeholders built the AIS with the goal of improving maritime safety, security, and MSA. The AIS was launched as a joint effort of the International Maritime Organization (IMO), the International Association of Marine Aids to Navigation and Lighthouse Authorities (IALA), the International Telecommunication Union (ITU), and the International Electrotechnical Commission (IEC) [1]. The IMO provides carriage requirements and performance standards. The IALA provides basic technical requirements, while the ITU and the IEC provide the broadcasting and electrotechnical standards of AIS. Although the development and deployment of AIS has contributed significantly to increased navigational safety, it has also resulted in numerous safety and security issues.

Originally, AIS was introduced on certain types of International Convention for Safety of Life at Sea (SOLAS) ships to assist the Officer of the Watch (OOW) in making decisions in the event of a collision. AIS provides the OOW with the most accurate and detailed data on all detected targets. This is very important for the OOW, especially targets that are at risk of or dangerously close to collision [2]. In addition, the data from AIS can be integrated into the Automatic Radar Plotting Aid (ARPA) and chart systems (Electronic Chart Display and Information System (ECDIS) or Electronic Chart Systems (ECS)) [2,3]. This provides the OOW with an additional source of navigation data (e.g., Closest Point of Approach (CPA)) that can help the OOW navigate or mislead him when the AIS station is spoofed. Although the use of AIS on ships has increased the safety of navigation, it is important to emphasize that AIS is not intended to guide navigation but to exchange navigational and other relevant data between ships and shore stations [4].

According to SOLAS regulation V/19.2.4, all ships of 300 gross tonnage (GT) and more engaged in international voyages and cargo ships of 500 GT and more not engaged in international voyages, as well as passenger ships, irrespective of their size, shall be equipped with AIS [5]. Behind these international regulations, the European Union and national authorities have also developed obligations regarding AIS for non-SOLAS ships (i.e., fishing and pleasure boats). To date, various AIS stations have been developed. According to IALA guideline 1082, AIS stations are grouped by class and function into mobile stations (shipborne class A and B, AIS SART, MOB-AIS, EPIRB-AIS, AIS and SAR aircraft) and fixed stations (AIS base stations, repeaters, and aids to navigation (AtoN)) [1]. Satellite-based AIS (S-AIS) has been introduced [7,8], extending AIS coverage from the VHF range up to 400 km [9], with diverse regional and global applications including maritime spatial planning [8,12].

AIS transponders transmit and receive specific information on dedicated VHF frequencies [13] using the self-organized time division multiple access (SOTDMA) protocol at 9600 bits per second with GMSK modulation. The ship transponder encodes the relevant information in the AIS message string to within 256 bits, and messages are fragmented using NMEA-0183 with a maximum sentence size of 82 characters. The protocol relies on a highly accurate time reference from the GNSS signal to which all stations synchronize to avoid overlap [16]. A single AIS channel carries 2250 messages per minute, doubled across two channels (87B at 161.975 MHz, 88B at 162.025 MHz). Typical range is 20 nautical miles (NM) ship-to-ship and up to 40 NM ship-to-shore, with VHF data link (VDL) congestion common in high-density areas — a context in which spoofing can endanger navigational safety.

Today's AIS data applications have shifted from collision avoidance, identification, and tracking to monitoring shipping routes, traffic trends, risk analysis, accident investigation, search and rescue, waterway planning, traffic simulation and forecasting, fisheries monitoring, and prevention of illegal activities at sea [1,8,16–21]. This paper focuses specifically on spoofing: a case study of an event near Elba in December 2019 confirms that typical maritime AIS can be easily spoofed to generate erroneous position information, and concludes by recommending a robust cybersecurity framework. The paper is structured as follows: Section 2 describes the methodology, Section 3 presents AIS vulnerabilities and practical examples of significant spoofing incidents, Section 4 a practical investigation of an AIS spoofing event, Section 5 the discussion, and Section 6 the conclusions.

2. Methodology

A systematic literature review followed by a comprehensive, explicit, reproducible method of data collection was conducted and structured following the documented guidelines [23–25]. This method consists of ten steps divided into three main phases. In the first phase, planning focused on defining a review question to guide the search: "What are the implications of AIS spoofing in the maritime domain?" The second (review) phase identified the appropriate time frame for documents from research databases such as Scopus, Science Direct, Web of Science, Google Scholar and open sources. "AIS" and "spoofing" were identified as search keywords. After refining the selection, documents were analyzed and synthesized by context, methodological approaches, and results. The final list included 55 documents limited to AIS spoofing (28 articles, 18 peer-reviewed journal articles, and 9 reports from specialized agencies) covering 2019–2020, with some earlier studies also considered.

In the third phase (reporting and dissemination), findings from the literature review are reported, corroborated by a case study analysis from the Faculty of Maritime Studies and Transport, University of Ljubljana, of a particular AIS spoofing event near Elba Island in late 2019. As part of the research, AIS data were first obtained through cooperation with the Slovenian Maritime Administration, stored at the MARES regional data exchange program. Additional day data were subsequently obtained from the Italian Coastguard and used to analyze signal strength received at the Elba AIS base station. The French agency maintaining the AIS system and commercial providers MarineTraffic and VesselFinder were also contacted. An investigation into the consequences followed; seven vessels were found to be in the spoofing cloud, and a letter was sent to those in charge of ISM implementation to enquire about possible consequences for shipping safety. Afterward, the archive data were streamed again using the AIS Network Data Client application and played in two different VTS applications, Navi-Harbor (Wärtsilä) and Pelagus (Elman). A navigation scenario was configured in the affected area, and the AIS spoofing data were displayed on ECDIS and RADAR applications via the Navi Trainer Pro ship tracking simulator. Finally, using the EMODnet method [26], a traffic density map (TDM) was created using ship positioning data from terrestrial and S-AIS sources, maritime infrastructure, and the SafeSeaNet Ecosystem Graphical Interface (SEG) application.

3. Results

This chapter highlights the unique AIS challenges associated with problems in securing ships at sea. It presents findings on AIS vulnerabilities, some recent examples of AIS spoofing trajectories, and practical investigations of an AIS spoofing event. To understand the current research, it is important to understand the background, how AIS works, the liabilities, and the methods by which vulnerabilities to the system can be initiated. The nature of the cyber arena is changing today, and the maritime sector is experiencing sophisticated and complex attacks that no one could have predicted years ago.

3.1. AIS Vulnerabilities

To date, some studies have been conducted on AIS vulnerabilities. Balduzzi et al. [27] discussed the security evaluation of AIS and divided the AIS threats into three macro-categories: spoofing, hijacking, and availability disruption, based on software or radio frequency (RF) threats. Kessler pointed out that the vulnerability of the AIS VHF broadcast frequency can lead to bandwidth usurpation by potential attackers who can prevent other devices from transmitting, negatively affect the synchronization process, or change slot reservation/assignment information [28]. Goudosis and Katskias [16], as well as Aziz et al. [9], emphasized that AIS-broadcasted messages are neither encrypted nor authenticated, making them vulnerable to unauthorized manipulations.

The vulnerabilities of the AIS arise mainly from its technical performance. The AIS is an open-source system that transmits on VHF channels in the maritime mobile band [1,3]. The data sent by the transponders are simply not subject to any control or verification of their integrity [16,29]. Data exchange between stations in the AIS service is largely based on transmission and reception at RF, and therefore the AIS service is vulnerable to malicious transmissions and risks of data manipulation [1]. At the center of the AIS service are base stations designed for use by government agencies to manage the VDL and provide wide Vessel Traffic System (VTS) or coastal surveillance coverage [1]. Base stations are vulnerable to possible attack if someone sends false AIS signals that can overload time slots and prevent or complicate the operation of regular AIS stations. AIS SAR devices (MOB-AIS, EPIRB-AIS, and AIS SART stations) transmit distress signals with geographic coordinates; if attackers generate false distress signals, this could deceive shipborne AIS stations, VTS, and SAR services. AtoN is also part of the AIS service that could enhance the safety of navigation, especially in coastal navigation, though not all ships are equipped with AIS and ECDIS [30].

From an operational point of view, AIS AtoNs are identified as physical, synthetic and virtual [1]. Based on the technical characteristics, it is obvious that synthetic and virtual AtoNs (V-AtoN) are more vulnerable to spoofing than physical AtoNs. The V-AtoN can easily be spoofed intentionally [31], and spoofed AtoN signals can mislead the OOW during navigation. Combining virtual and physical AtoNs is considered an effective measure to avoid the negative consequences of V-AtoN spoofing [32]. The possibility of V-AtoN spoofing should be considered the main vulnerability given AIS dependence on GNSS, which is also vulnerable to spoofing [30,33]. Ship-borne Class A and B stations, as the main component of the AIS service, could enhance MSA and collision avoidance, but AIS vulnerabilities can result in different errors and falsifications of messages with significant impact on mariners and local authorities: a malicious attack could produce a false maritime picture misleading the OOW or VTS and affecting navigational decision-making.

Iphar et al. [34] presented a risk analysis of the AIS supported by the EBIOS methodology and an anomaly typology. The EBIOS method consists of a 5-module procedure: context study, study of dread events, study of threat scenarios, study of risks, and study of security measures. Following on from the EBIOS risk analyses, Iphar [34,35] presented different dread and threat scenarios and risks: 11 risks with an intolerable level linked to AIS use were identified, concerning integrity and availability — dynamic data, AIS data, transmission and reception functions, and positioning and visualization functions. Some identified risks deal with data integrity issues, one of the data quality dimensions; different kinds of possible anomalies in AIS messages were recognized (see Table 3 of the original paper). To assess AIS data and detect anomalies, different algorithms and filters can be used to provide the maritime surveillance community with information on unusual behavior [36]. To prevent AIS security weaknesses, several solutions have been proposed: protected AIS (pAIS) using public-key cryptography for sender authentication and message integrity [28]; the secure AIS application protocol based on pairwise keys with a certification mechanism [9]; Auth-AIS using the TESLA-based Deterministic or Probabilistic (with Bloom filter) Security Configuration models [15]; and the maritime certificate-less identity-based public-key cryptography (mIBC) architecture [16]. All these proposals are cryptography-based, backward compatible with existing AIS devices.

3.2. An Overview of AIS Spoofing Incidents in the Last Years

The last two years of the decade have been remarkable in many ways. While the coronavirus pandemic disrupted operations worldwide, some global trends continued unabated — including our increasing reliance on GNSS [38,39]. In recent years, several disruptive incidents caused a stir in the shipping industry (see Table 4 of the original paper): identity fraud for illegal fishing in the Southern Ocean (2008–2018); m/v Ramtin falsely transmitting her identity off Oman/Malaysia (2013); 9,883 suspected spoofing incidents at ten global locations (2016–2019); the Black Sea incident near Novorossiysk (June 2017) placing vessels at Gelendzhik Airport; m/v Yuk Tung spoofing in the East China Sea (2018); mysterious GPS ring patterns off Point Reyes, California (2018–2019); Eastern Mediterranean and Red Sea interference (2018–2019); seizure of the Stena Impero after being spoofed into Iranian waters (July 2019); m/v Fu Xing 12 manipulating its identity with four ship names (2019); fake ring-pattern signals at the Port of Shanghai (2018–2019); four fake visual AtoNs at Ponce De Leon Inlet, Florida (2020); the deliberate Elba spoofing (3 December 2019); and a fishing fleet misreporting its location ~10,000 km from its observed Galápagos position (July 2020).

From these cases it can be concluded that the original purpose of AIS spoofing to this day is to disguise illegal fishing and other illegal activities at sea, including ship spoofing and AIS hijacking, with some examples of AtoN spoofing (Florida). In recent years, GNSS spoofing has appeared as part of defense development in civilian scenarios, and AIS spoofing has been deliberately used for electronic warfare and concealment of military activities, as in the Eastern Mediterranean and the Red Sea. The Black Sea incident led to speculation it could be attributed to satellite spoofing technology tests by one of the space superpower states — as part of their electronic warfare arsenal or simply as an anti-drone measure to protect very important persons [40]. A practical example of geopolitical and geo-economic competition is the July 2019 detention of the British oil tanker Stena Impero, seized by Iranian forces after she was tricked (spoofed) into altering course into Iranian waters, making ship, cargo, and crew pawns in a geopolitical war [41]. Many shipping companies in the region instructed their ships to transit Hormuz only at high speed and during daylight — while one-third of the world's seaborne oil, about 17 million barrels per day, flows through the strait [40,41].

At the same time, a mysterious new electronic device emerged in China that spoofs AIS signals in ways experts have never seen: multiple spoofing incidents across 20+ coastal areas and ports (Shanghai, Fuzhou, Qingdao, Quanzhou, Dalian, Tianjin) ongoing for months. Unlike 'traditional' spoofing, GNSS signals were congregated into large circles, later dubbed "crop circles," with moving signals shifted to the same position, creating a confusing traffic situation for pilots [42]. Bergman [43] observed the "spoofing circles" were at oil terminals, and the timing — US sanctions on Iranian oil purchases — plus observations of Iranian oil entering China suggest some spoofing disguises these transactions. Circulation phenomena were also observed in Tehran [44]; in all these cases the real location was relatively close to the fake circling locations. In other 2019 cases, however, ships' real locations were thousands of miles away — literally on the other side of the globe — with true positions verified via the field of view of the receiving satellite. It remains unclear whether these errors were caused by the ships' AIS or by error/interference with GPS receivers.

Another example was observed in July 2020, when one of the world's top fishing fleets was accused of misreporting its location to conceal illegal fishing in the EEZ around the Galápagos Islands: vessels reported via AIS a location ~10,000 km from their observed position while actually near the Galápagos, where illegal fishing had occurred before. The HawkEye360 satellite system — providing maritime awareness through RF geospatial intelligence — revealed many instances of suspicious behavior, including evidence that many vessels stopped transmitting AIS for eight+ hours and may have penetrated deep into the Galápagos EEZ [38,45]. This type of AIS "cloaking" is just one of many ways criminals use location spoofing of a GNSS-dependent system [38]. The questions of the Florida AtoN manipulation and the late-2019 Elba AIS spoofing remain open; the investigation of the latter is presented in the next section.

4. Case Study on Spoofing Large Amounts of AIS Fake Targets Nearby Elba Island and Its Impact on VDL

With increasing demand for maritime VHF data communications, AIS is heavily used for maritime security, situational awareness and port security [46]. VDL exposure remains an increasingly serious problem worldwide due to proliferation of AIS applications, message types, services, and equipment types, plus unanticipated increases in user traffic. The critical 50% VDL threshold has already been exceeded in some high-shipping areas. However, when intentional emission of virtual targets occurs on top of high AIS traffic, such actions can further jeopardize system operation. IALA has indicated additional AIS channels are needed to provide maritime safety information and general data communications and to protect VDL integrity [46].

Each ship transmits its AIS data and receives AIS data from all ships within VHF range. In high traffic density, it becomes necessary to reduce the number of vessels in a communication cell; when AIS messages begin to overload the network, the AIS station can automatically shrink its cell by ignoring weaker, distant stations in favor of nearby ones. When the data link is congested to the point that safety information transmission is compromised, two methods resolve congestion [47–49]: first, intentional slot reuse from the most distant station(s) within the candidate slot selection interval (the "Robin Hood" principle [50]); second, base-station allocation of Class A reporting rates to non-Class A stations and redirection of slots to reserved FATDMA slots, allowing the broadcast system to be overloaded by up to 500% while still providing close to 100% throughput for ships closer than 8–10 NM in ship-to-ship mode. In practice, system capacity is almost unlimited.

The only known AIS spoofing case with a large number of fake AIS messages occurred on 3 December 2019, starting at 13:13 UTC, between Elba and Corsica. Thousands of AIS streams were received and recognized as Dutch-flagged naval units, artificially generated with different identification codes, positions, routes, and speeds, completely saturating a circular sea area of ~11 NM radius. The attack was repeated three times: 13:13 UTC (3 min), 13:28 (4 min), and a final burst at 13:37 lasting only seconds. During the first burst, 2,389 messages were received, all AIS message No. 1, with MMSI numbers from 244800009 to 244803507. In the second burst, MMSIs ran from 244803510 to 244807010 — presumably the spoofer transmitted all consecutive MMSI numbers, 3,500 in each broadcast. The third broadcast was aborted immediately for unknown reasons. A total of 5,241 AIS messages (3,741 distinct MMSIs/ships) were received. The second transmission also carried many AIS message No. 5 reports with ship identification and dimensions: all ships identified as passenger ships (AIS type 60), 90 m long, 24 m wide, no draft or destination. Vessel names and call signs were generic — VESS0 and CAL plus the last three MMSI digits (000–999).

The original paper's figures show the spatial distribution of received messages (1st and 2nd bursts): real targets, a cloud of artificial targets, and marked spoofed targets of interest. At the very start of the first spoofing, two notable targets appeared at the cloud's edge: one with invalid MMSI 999999999 — common around AIS and often associated with navy vessels — and a Belize-registered bulk carrier (MMSI 312320000) scrapped since 2016 (fishing vessels are known to assume scrapped vessels' identities). At the start of the second spoofing, notable ships again appeared at the cloud's edge: a generic artificial ship emerged from the cloud very close to the first burst's ships, and one ship appeared as a coastal station with invalid MMSI 4803509 — generated by truncating the first two digits of the artificial ship's generic MMSI data (MID 24480****), i.e., probably 244803509 converted into an AIS base-station signal. A fake US-registered vessel (MMSI 367309390) also appeared on the south side. Among the artificial targets, MMSI 244805000 stood out — a fishing vessel number spoofed several times in the past — placed directly in front of the real vessel Crudemed, transmitting static data like many in the second burst, at 40 knots.

4.1. VDL Load and Spoofing Source Analysis

First, ITCG provided received data from other vessels (VDM) showing most AIS messages were received via the Monte Capanne AIS base station on Elba Island; only 20 artificial-ship messages were received by the Corsica base station. Real-ship signals were almost all received from other base stations such as Monte Agentario or La Castellana. The MMSI range used in spoofing along with repetition counts (1–4) per MMSI — and missing MMSIs — indicate a spoofing algorithm running with auto-incrementing MMSI numbers; missing MMSIs are likely due to transmission collisions with nearby vessels.

The distribution of spoofed targets during the second burst — principal component analysis with two principal axes, ship positions in UTM coordinates — appears most likely a parabolic spatial distribution, while ship velocities and courses are normally distributed (model in Equation (4) of the original paper). Gaussian curves along the primary axis for both bursts show sources shifted by ~5 km with the same standard deviation of 7.7 km. The EMODnet method showed shipping density up to 45 ships/km² [26]. Regarding VDL load, the number of slots with RSSI above noise indicates the spoofer generated almost 100% VDL load. AIS message 4 (base-station report) shows the base-station memory saturated — Monte Capanne received more than 2,048 different AIS station reports in the last 9 minutes.

Finding the AIS generator position is difficult because signals were not transmitted synchronously with the AIS frame, so time-of-arrival (TOA) information cannot be relied on for ranging [51]. The first candidate is the vessel with MMSI 999999999; its trajectories over several months show it most frequently in and off the main naval base in the area. It is very likely the naval-base ship differs from the one received near Elba, which may in turn be the same one received again near Bonifacio hours later — suggesting several naval vessels sharing the same artificially generated MMSI, not unusual for warship AIS devices. Warship systems can generate such a spoofing event. MMSI 312320000 and MMSI 367309390 are also suspected: notably, the jitter error and RSSI levels of these two targets match the Dutch-MMSI spoofed messages and MMSI 999999999 — possibly indicating a naval ship conducting an electronic warfare exercise using two MMSIs. However, two transmitters sending AIS at the same power can be received with the same RSSI even from entirely different positions (equidistant from the receiver). Still, warships in such activities likely wish to keep a low profile. Fishing vessels cannot be ruled out, but they usually prefer to avoid attention and use simpler AIS manipulations — though they are becoming increasingly adept.

Received-signal analysis shows rapid changes between transmissions, possibly indicating the transmitting and receiving stations are not in line of sight and subject to signal fading (a ship rolling on waves may contribute). A large abrupt drop at the end of the first burst is not fully understood — possibly a blind sector or obstruction. A slight constant decrease in average RSSI from mid- to end of the second burst could indicate increasing distance between transmitter and receiver — again pointing to a moving spoofer.

4.2. Potential Consequences of Spoofing Near Elba Island

Seven vessels were found in the spoofing cloud (see Table 5 of the original paper: La Suprema, Blue Star, Eliana Marino, Mega Express Five, Leale, Crudemed, Cruise Ausonia). A letter was sent to those in charge of ISM implementation on those vessels to enquire about possible consequences for shipping safety. Even without their response, a navigation scenario was configured in the affected area: archive data were streamed using AIS Network Data Client and played in two VTS applications (Navi-Harbor by Wärtsilä and Pelagus by Elman), with spoofing data displayed on ECDIS and RADAR via the Navi Trainer Pro ship tracking simulator.

For this case study, m/v "Eliana Marino" within the spoofing-cloud area was selected. The clock was synchronized with AIS data, and using her actual position, course and speed, what appeared on her ECDIS screen just a minute before the spoofing was simulated: navigation data were appropriate given matching ENC, radar, and AIS data — allowing proper MSA. The first spoofing burst displayed on ECDIS with radar overlay and AIS data then caused significant MSA degradation, with 95% AIS signal processing load. Consequently, the radar image showed m/v "Eliana Marino" on a collision course with more than a dozen fake vessels. Filtering sleeping AIS targets and limiting ships to a 6 NM range made the picture slightly clearer, relieving the ECDIS/ARPA system, with CPA and TCPA filters adjustable — yet from the OOWs' perspective the situation remains hazardous: many false collision-course signals with very short CPA generate many collision alarms, potentially leading to inadequate decisions. An experienced OOW will use a "raw" radar picture without AIS support and sharpen visual look-out. Luckily the event occurred in daytime in a favorable area; at night in dense, dangerous traffic it could have severe consequences. OOWs must therefore be aware of possible AIS spoofing and its impact on MSA. VTS overload was confirmed on both tested applications (Navi-Harbor and Pelagus).

5. Discussions

Along with Iphar [35] and several other studies, this research has demonstrated AIS shortcomings: weakly secured and vulnerable to erroneous information, data corruption, and data spoofing. Before this Elba December 2019 case study, only a few cases — the Black Sea [38], the Strait of Hormuz [40], and the Port of Shanghai [42] — were known to noticeably impact shipping.

In the Elba case study, the lesson internalized is that a spoofing situation generating many false collision-course signals with very short CPA can severely endanger navigational safety. The root cause of successful AIS spoofing is not only that even the cheapest online jammers can cause complete receiver outages [52], but a lack of appropriate technical security controls. An AIS user may mistakenly alter course to avoid a non-existent ship — severe in dense shipping areas or near natural obstacles. Here the system also reached maximum performance, revealing AIS saturation from high user numbers. Paradoxically, this problem may grow as IMO [53] increasingly equips more ships with communications technology. From this perspective, the VHF Data Exchange System (VDES) is seen as effective, efficient spectrum use building on AIS capabilities [54]. Unlike AIS, VDES is far more robust and could benefit from authentication mechanisms such as digital (encrypted) signatures to discard spoofed messages.

As noted, AIS devices no longer follow today's technological developments and cyber threats. Manufacturers should offer devices providing reliability and resilience validated by the most demanding systematic deployment methods an attacker would use — penetration testing [55], performed by qualified testers using techniques such as simulated risk assessments, attack scenarios, remote login solution security, social engineering, and testing of physical perimeters protecting computer and network devices. Testing should preferably target production systems against close-duplicate test systems, or pre-production systems such as mandated acceptance testing. Testers must assess how AIS receivers respond to spoof signals and use this to develop counterstrategies increasing interference resistance. A set of standard tests should let AIS users select the best equipment for their application based on protection against jamming and spoofing — achievable only through closer collaboration between the maritime sector, government security organizations, industry, and academia.

In other words, security requirements can be met with robust measures accounting for currently known threats during production of new devices and systems. Yet the history of technology suggests a sober review also requires imagination: no sane historian of nuclear weaponry would advise any technology stakeholder that their technology has reached a stable, attack-proof point. Techniques live lives of their own. Common sense dictates banning nuclear devices, yet they are produced and sold to super-redundancy — not all a matter of profiteering, since techniques also capture the imagination of great scientific minds, with missile defense systems inviting attacks that prevent non-scientific processes from eliminating such weapons. Hence technique devours the unstable geopolitical and economic world. Rather than resolving human problems through human means, humans perpetuate advantageous positions through technological advantages most cannot afford — raising serious ethical questions about what some call technology abuse (wealthy nations) and others mere opportunistic use of available techniques to meet oppressors on equal footing. On this topic, there is little reason to believe securing trade will ever relax; realistically, expect the cycle of security, insecurity, and demand for technical advances to continue.

6. Conclusions

AIS represents a valuable tool for building MSA and enhancing navigational safety. It has numerous vulnerabilities and pitfalls as an open-source system transmitting on dedicated VHF frequencies — simultaneously the system's main advantage and disadvantage. The simplicity of the AIS protocol has led to today's wide variety of applications, while AIS still lacks message integrity verification and authentication capabilities. From the navigators' point of view, verifying AIS information using radar and visual look-out is of great importance. The essential navigational safety problem is AtoN spoofing, especially V-AtoNs, whose message integrity is not easy to verify — navigators should pay special attention and use all available sources to verify V-AtoN messages. Experienced navigators can easily detect some AIS spoofing types, such as abnormal target trajectories or large numbers of spoofed signals, by comparing with other sources. Navigators should never rely on a single information source and should double-check AIS-provided data.

A serious geopolitical consideration is that in an unequal, fiercely competitive economic world, certain nations — when justifiably finding various maritime interference unfair — will use vast resources to bypass current protocols. Scientists working to make AIS spoof-free face such countries and allies, who feel no compunction about violating rules and sanctions they consider blatantly unjust and inequitable, devoting vast resources to evading control mechanisms. In summary, the importance of cybersecurity has been highlighted: GNSS spoofing has been a defense issue for years and is now affecting shipping. As more devices and autonomous systems rely on GNSS, even more systems could be vulnerable to spoofing. The maritime industry and shipping are not immune to such cyberattacks, nor is the situation expected to improve soon. With hackers constantly seeking new ways to spoof and exploit AIS vulnerabilities, new cybersecurity openings will emerge unless systems are adequately protected. This analysis showed that even a relatively easy-to-obtain AIS spoofing generator, such as that near Elba Island, can impact ship security: the mass of ships on screen is primarily a technical problem projecting a clearly false scenario, under which a real ship can be missed — making simultaneous use of other safe-navigation means essential. When both AIS and GNSS, on which precise positioning is based, are subject to spoofing and denial-of-service attacks threatening to render AtoN useless or even endanger navigation, relying solely on ECDIS and its overlays may be unsafe.

Given the impact of digital technologies and to maintain seaworthiness, robust cybersecurity frameworks are needed. Disruptions to GNSS-based positioning and navigation have become global; to address a global problem, the GNSS community needs a global solution. The maritime industry must stay ahead: manufacturers need to ensure reliability, resilience, and functioning of multi-sensor systems for safety and liability reasons. Penetration testing systematically employing attacker methods is recommended, testing systems and sensors against close-duplicate test systems or pre-production systems [55]. GNSS signals are essential for safe, efficient navigation and integral to maritime navigation; their degradation jeopardizes maritime safety. This paper identifies AIS vulnerabilities impacting maritime security and recommends the maritime community implement a robust cybersecurity framework, use multiple GNSS constellations and encrypted signals increasing protection against spoofing and other maritime cyber threats. GNSS is not easily vulnerable — but beware those who say the risks are overstated.