feel it with your fingers / feel it with your toes
AI is all around me/And the feeling grows
(Apologies to Reg Presley)
Artificial intelligence (AI) has quietly played a key role in many aspects of maritime transportation systems (MTS) over the past few years. In one form or another, AI has been around for more than 70 years, but it wasn’t widely reported in public media until about a decade ago, and it finally reached critical mass in terms of public awareness last year. It seems that I did. In 2023, there will be many articles about AI in various aspects of life, including articles related to the Hollywood actors and writers’ strike, and there will be many questions about what AI is and how humanity Everyone seems to have an opinion on whether it’s progress or not. The birth of the devil and how it should be regulated and legislated.
AI refers to the integrated intelligence of machines and software. They have been anthropomorphized primarily in the form of humanoid robots in science fiction, such as Isaac Asimov’s Robot Stories (1940-1955) and Robbie the Robot in the 1956 classic film Forbidden Planet. AI became a serious academic research field in the mid-1950s, and its format has changed over time to reflect changes in computer and peripheral technology. It is no exaggeration to say that AI research began in 1950 with Alan Turing’s description of his imitation game. Turing His test is an indicator of whether humans can interact with machines that are indistinguishable from interactions with other humans. There is some debate as to whether Chat Generative Pre-trained Transformer (ChatGPT) has passed the Turing Test, but if it hasn’t passed it yet, it may be very close to passing it.
Modern AI (since 2010) relies on deep learning, a combination of machine learning, big data, and neural networks. The digital revolution that began in the 1960s was key to the development of big data. By digitizing all forms of data and communication, we can now build vast data sets that can be searched, analyzed, and transformed in limitless ways to become the learning database for AI systems. A neural network refers to a computer that can learn from collective knowledge distributed within the network, rather than relying solely on its own programming. This was first demonstrated in his 1952 by the Stochastic Neural Analog Reinforcement Computer (SNARC). Today’s AI research goals include advances in knowledge representation, reasoning, planning, natural language processing, and perception to build machines that demonstrate general intelligence and the ability to solve arbitrary problems. . AI is already playing an increasingly important role in all of our daily lives, but that role is often less obvious or invisible. Common applications include advanced search engines, marketing and “recommendation” systems, human voice recognition, self-driving cars, authoring/art tools, strategic and serious games, and medical diagnostic and treatment tools. .
As an example of how AI is advancing, a British computer scientist filed a patent application with the Intellectual Property Office (IPO) on behalf of an AI he had invented. The IPO rejected these applications in 2019, stating that only individuals can issue patents. Although the inventors argued that AI is a “form of conscious and sentient machine intelligence,” the UK Supreme Court upheld the IPO’s ruling in December 2023.
Although these concepts seem far removed from maritime and cybersecurity, AI is already an integral part of both. Perhaps there is no need to consider all the ways in which automation and autonomy in maritime are being enhanced to some degree by AI, from smart ships, ports and containers to fully autonomous navigation and ships. However, automation and some autonomy should not be confused with AI. Many automated systems read a series of sensors, dials, and other inputs. If the system state is X, the action is Y. This is purely algorithmic. It’s not intelligence per se, it’s programming.
Interactions between AI and cybersecurity fall into three basic categories, often referred to as defensive, offensive, and adversarial. As the name suggests, defensive AI refers to techniques used to aid in cyber defense. AI can help detect cyber fraud, anomalous email messages and data traffic patterns, and phishing attempts. Intrusion detection and prevention, log analysis, incident response and recovery strategies and procedures can be aided by the use of AI tools. Analyzes from AI tools can help mitigate denial of service (DoS) attacks and even predict potential software vulnerabilities and zero-day exploits. AI-driven risk management planning and patch management can be much more efficient and optimized than manually managing these processes.
Offensive AI refers to counter-defensive techniques used to help plan and execute cyber attacks. AI can collect information from the internet to quickly generate highly personal and effective spear-phishing messages and other forms of social engineering attacks. Given AI’s ability to craft well-written, plausible messages that sound accurate and conclusive, it seems like misinformation and disinformation is almost AI’s specialty. AI does simple tasks like making deep fakes that look legitimate, data poisoning, and data traffic manipulation. AI tools make password cracking, automated hacking, and botnet management much easier.
The third form is adversarial AI (AAI), which is a method of directly attacking other AI systems. AAI methods degrade, deny, deceive, or manipulate AI systems. Attack models used by other AI systems, add noise to a system to confuse an adversarial AI, reprogram or introduce errors into other AI software, or modify training data used by an AI system. There are many adversarial techniques that can be used, such as poisoning. Other AI systems.
Maritime transportation systems have some unique cybersecurity needs due to the unique environment in which we operate and our unique information and operational technology systems. AI will become part of the cyber products, procedures, and strategies we employ for the same reasons as other industries. AI will be able to detect, predict, and test potential vulnerabilities and attack schemes before anyone else. AI will be able to create and even predict attack and defense mechanisms faster than humans. It can also detect and predict errors reported by the Global Positioning System (GPS) and Automatic Identification System (AIS) and estimate the tracks of other vessels in real time. And as AI learns about maritime systems in record time, the learning curve for both offense and defense for MTS will be steep.
We encourage all MTS stakeholders to learn how and where AI fits into their organization, systems, and requirements. There is no doubt that we all need AI for cyber defense, but we cannot rely on it alone. We still need to focus on building and maintaining a culture of cyber safety, practicing basic cyber 101 well, and focusing on finding and mitigating vulnerabilities.
I’ll repeat what I’ve been saying for years. “Those who think technology can solve their problems don’t understand technology or their problems.” Understand the role of AI and understand (and yet another) technology, especially (really!) It is essential to avoid over-reliance on technology that is not available. AI will be the basis for many tools that support his MTS employees, but it is far from a replacement.
Dr. Gary C. Kessler, CISSP, is the principal consultant for Fathom5. This article is excerpted and expanded from Maritime Cyber security: A Guide for Leaders and Managers by Kessler and Steve Shepard.
The opinions expressed here are those of the author and not necessarily those of The Maritime Executive.