The AI-ARC project applies AI, machine learning and VR technology to provide a “smart operational picture” of the increasing marine traffic to the Artic Sea, by automatically filtering “numerous validated and statistical data streams and databases to a user-friendly interface”.
The resulting AI-based platform, called “the Virtual Control Room” will allow “users to specify their preferences in terms of threat levels, abnormal behavior, interoperability and risk management by flagging detected anomalies with confidence and providing threat or risk levels according to a predefined model based on user preferences”.
According to the project Consortium, this will permit strengthened maritime “situational awareness” and better decision-making without increasing the workload on its users. This “predictive artificial intelligence” will be used to predict icepack and iceberg movement, vessel traffic and “to predict the drift track of a missing vessel” and more generally improve coordination in Search and Rescue operations.
Artificial Intelligence, machine learning, and virtual reality are all part of AI-ARC’s technological arsenal.
AI-based services provided by project outputs include “Ice Pack Prediction, Vessel Traffic Prediction Service, Risk Index Computation, Reliability Assessment of ML-Services, Satellite Based Vessel/Ice Detection Service, Satellite-Based Sea Coverage Ice Maps Service, Anomaly Detection in Vessel Behaviour, Detection of Ice Block and Trajectory Service, DBN Anomaly Detection Service, SADV Anomaly Detection Services, and Anomaly Detection Earth Satellite Data” (listed and detailed in D4.5).
“Data Acquisition and Fusion Services” are also described in the AI-ARC Service Catalog, to merge and verify “data from multiple sources – from vessels, actions, locations, agents or objects – to support end user services such as behaviour prediction, risk assessment, anomaly detection, visualization of analytics and decision making.”
The project’s Cordis page writes that “The AI-ARC team is currently working to develop its solutions towards commercialisation in the VIGIMARE project starting September 2024.”
VIGIMARE, or ‘Vigilant Maritime Surveillance of Critical Submarine Infrastructure’, is a three-year Horizon Europe project which aims at providing “an information sharing environment to the Critical Infrastructure Owners to build their resilience against physical, cyber and hybrid threats to the EU submarine Critical Infrastructure.”
According to a deliverable from the NESTOR project (D7.6), AI-ARC is also involved in the creation of a “preliminary standardisation roadmap for the border management domain” together with the NESTOR, BORDERUAS, EFFECTOR, ENTRANCE, ISOLA, MEDEA, PROMENADE, ROBORDER, and METICOS EU-funded projects.
Lastly, AI-ARC’s outcomes could inform and provide collaboration opportunities with the EURMARS project, writes EURMARS deliverable D6.1.
Project outputs were tested in two demonstrations:
1) The “Baltic Demonstration”, which took place on the 21st of September 2023 in Karlskrona, Sweden. There, the Digital Map Table (DigLT) feature — a software system for shared situation visualization and analysis — was demoed in several different use case scenarios (such as for environmental protection, the protection of critical infrastructures, and to contrast illegal fishing and smuggling);
2) The “Arctic Demonstration” — which took place in Reykjavik, Iceland, on the 19th of October 2023 — to demo some additional features of the DigLT software, including a “first live use case” providing — among other data — “the search and rescue area (SAR) of Iceland and some key SAR assets”, and AI-based vessel track prediction. A “3D VR-based application” was also used for DigLT “to review the information in a true 3D context”.
Deliverable D4.4 provides a detailed overview of both demonstrations, concluding that “the project goals have been achieved”. As a result, “the VCR (Virtual Control Room) has a strong potential to evolve into a commercial product. Therefore, future projects should explore the diverse applications of VCR and raise its current TRL (Technology Readiness Level) in the services that have already been researched and developed in the AI-ARC project”, concludes the deliverable.
The project seems to be informed by a typical solutionist premise according to which “Artificial Intelligence needs to be taken promptly into account by the maritime operators and search and rescue agencies in order to be step ahead rather than late in collecting relevant data at right time at right place” (Policy Brief #2). As a result, one might wonder whether ethical considerations have been taken into account as radically as needed.
Results detailed in D1.11, the ‘Societal Impact Assessment Report’ seem to reflect a similar concern: “The main revelations of the Societal Impact Assessment (SIA) for the AI-ARC project are that that the maritime experts highlighted a very large variety of different the ethical challenges and potential negative impacts of increased safety and security in the Arctic, including privacy concerns, resource exploitation, and SAR operations. However, they saw equally that the solutions proposed included transparency, common regulations, and limiting alerts, which (hopefully) result in positive outcomes, along of course the obvious attempted outcome to the society: enhancing maritime safety and security – ultimately saving lives. Also, possible outcomes include innovations that lead into environmental conservation, and improved situational awareness. The remarks analysed and gathered in this report emphasise a balanced approach to addressing ethical concerns while promoting sustainable development in the Arctic.”
The fact that “positive outcomes” can only be hoped for is clearly indicative of the above-mentioned concerns, which also include “the risk of privacy infringement as security efforts intensify”, the fact that “The introduction of advanced surveillance systems and other security measures raises alarms over potential government overreach and abuse of power”, questions over “whether societies are willing to accept similar levels of surveillance as seen globally, for the sake of safety”, the “Sustainability and ethical implications of maintaining data centres for processing maritime data”, and even the importance of addressing “concerns of plagiarism and originality by acknowledging the iterative nature of human and AI creativity”.
