Artificial intelligence is rapidly becoming another everyday business tool.
Marine businesses are already using systems such as ChatGPT and other AI platforms to help write emails and product descriptions, translate material, prepare marketing content, generate images, answer customer enquiries and analyse information.
For businesses with connections to Europe, however, the use of AI is increasingly becoming a compliance issue as well as a productivity opportunity.
New transparency requirements under the European Union’s AI Act took effect on 2 August 2026, adding to AI literacy requirements that have applied since February 2025. Importantly for the international marine industry, the legislation is not limited simply to companies headquartered within the European Union. Depending on how AI is being used, businesses outside Europe may also fall within its scope where their AI systems or outputs are used in the EU.
That could potentially include Australian boatbuilders, equipment suppliers, yacht managers, brokers, charter businesses and other marine companies. What matters is not simply having European customers, but whether the output of the AI ends up being used in Europe. If a boatbuilder uses AI to draft a specification for a European yard, if a broker uses it to produce listing copy read by European buyers, or if a yacht manager runs an AI assistant that European crew and owners deal with, that output is being used in the EU.
Some readers will have seen reports that the EU has delayed the AI Act. It has, but not this part. The Digital Omnibus, which came into force on 27 July 2026, pushed the rules for high-risk AI systems back to December 2027 and August 2028. The transparency rules and the AI literacy duty were not delayed.
Transparency becomes important
One of the central principles of the new rules, set out in Article 50 of the AI Act, is straightforward: people should be able to recognise certain circumstances in which they are dealing with AI or viewing content created or manipulated by it.
For example, providers of AI systems that communicate directly with people, such as AI chatbots and virtual assistants, must ensure users are informed they are interacting with AI unless that fact is already obvious.
There are also requirements covering AI-generated and manipulated content.
This does not mean that every photograph adjusted using AI or every piece of marketing copy prepared with the assistance of ChatGPT suddenly requires a large “Made by AI” label.
The rules are more specific.
Providers of generative AI systems have obligations to ensure synthetic text, images, audio and video can be identified in machine-readable form. Businesses using AI systems also have disclosure obligations in particular circumstances, including the use of deepfakes and some AI-generated material dealing with matters of public interest where there has been no meaningful human review or editorial control.
The distinction is important because the EU rules treat the company developing or supplying an AI system differently from the business simply using one.
If your website runs a chatbot, the obligation to build in the disclosure that tells people they are dealing with AI sits with the company that supplies the system. Ask your vendor to confirm in writing that it does. Then look at your own site and check that the disclosure actually appears where the conversation starts, because you are the one whose name is on it.
For the average marine business, the immediate lesson is not that every use of AI needs a warning label. It is that businesses need to know where AI is being used, what it is producing and whether customers or the public are interacting directly with it.
There is one piece of breathing room. Generative AI systems that were already on the market before 2 August 2026 have until 2 December 2026 to meet the machine-readable marking requirement. Systems launched after 2 August 2026 had to comply from the first day.
Breaches of the Article 50 transparency requirements can attract fines of up to €15 million or three per cent of worldwide annual turnover, whichever is higher. For small and medium businesses the calculation is reversed, and the fine is capped at whichever of those two figures is lower.
Staff also need to understand the tools they use
The second issue is arguably more relevant to the everyday marine business.
Since 2 February 2025, the AI Act has required businesses covered by the legislation to make sure that staff and other people using AI on their behalf have a sufficient level of AI literacy. In July 2026 the wording was softened. Businesses must now take measures to support the development of AI literacy, and the amendment says expressly that no business has to guarantee any particular level of ability in any individual. The duty is about the effort made, not the result achieved. It still applies to every business in scope.
In plain English, businesses should not simply give employees access to AI and assume they know how to use it safely.
A salesperson using AI to prepare quotations has different risks from a marketing employee creating promotional material. Someone preparing technical documentation has different responsibilities again.
Staff need enough understanding to recognise what the technology can and cannot reliably do, when its output needs checking, what information should not be entered into it, and when a human needs to make the final decision.
This does not necessarily mean sending every employee on an expensive AI course.
The European Commission says there is no requirement for a particular certificate. Businesses can determine appropriate training or guidance according to their circumstances and can keep their own internal records of what has been provided.
For a small marine company, that may be as straightforward as identifying which AI systems are being used, establishing some basic rules around their use, making sure employees understand those rules and recording that the discussion or training has taken place.
Why keeping a record matters
The paperwork may ultimately be almost as important as the training itself.
Marine businesses are already accustomed to documenting procedures, inductions, qualifications, safety training and maintenance. AI governance is likely to develop along similar lines.
If an AI-generated document causes a dispute, confidential information is inadvertently disclosed, inaccurate technical information reaches a customer or a misleading image is published, one of the obvious questions will be how the business controlled the use of the technology.
Being able to demonstrate that the company had rules in place, that staff knew those rules and that somebody was responsible for checking important AI-generated work provides a very different position from simply saying that employees were expected to know what they were doing.
Insurance companies, customers, regulators and lawyers are also increasingly likely to ask businesses how AI is being used and what controls are in place.
Start with some simple questions
For many marine businesses, complying with the changing environment does not need to begin with a complicated technology project.
It begins with knowing what is already happening inside the business.
Does anyone use ChatGPT, Copilot, Gemini or another AI service at work? Is AI being used to produce quotations, contracts, product information or technical documents? Does the website use an automated chatbot? Are AI-generated photographs, videos or voices being used in marketing? Are employees putting customer or commercially sensitive information into public AI tools?
Once a business knows the answers, it can decide where some simple rules, training and record keeping are required.
The European Commission’s latest guidance makes clear that AI literacy does not require a prescribed training course or a particular governance structure. What is appropriate will depend on the business, the people involved and how the technology is being used.
For marine companies with European operations, customers or other exposure, however, doing nothing is becoming increasingly difficult to justify.
AI may be new technology, but the business principle behind the regulation is familiar: know what tools your people are using, understand the risks, establish sensible procedures and be able to demonstrate what you have done.
What to know more? Watch this video.
Kristina Agustin
Kristina Agustin is the founder of Southern Sky AI, where she works with marine and maritime businesses on AI governance, AI deployment and AI training. She is an Admitted Lawyer (Supreme Court of NSW), is completing a Master of Artificial Intelligence at the University of New England, and holds certifications including AWS Certified AI Practitioner, IWAI Certified AI Consultant and CPD Certified AI Trainer. She is a 2026 ATSE Elevate Scholar, and Southern Sky AI is listed on the Australian Government’s National AI Centre directory.
Before AI, Kristina spent more than 20 years in international superyacht and maritime operations. She started at sea, moved into senior interior roles on 50 metre-plus vessels, and then into yacht management, where she handled contracts, compliance and risk for major fleets. That is where the judgment comes from. She knows what a marine business can carry and what it cannot.
Her work covers governing the AI already inside an operation, deploying the tools it needs, and training crew and shore teams to a standard the business can point to. She writes The Chart Room Dispatch, a weekly email on AI and maritime, every Sunday at southernsky.ai/dispatch. Connect with Kristina through LinkedIn.
