As artificial intelligence progresses at breakneck speed, in what ways is it being used to disrupt the field of yacht design, and do we need to proceed with caution? Risa Merl looks at how AI is already making its presence felt.
As I gaze at the computer screen, a yacht emerges before me. It’s like nothing I’ve ever seen on the water – a fairly typical hull, painted turquoise, married to a superstructure resembling a towering Egyptian pyramid adorned in gilded patterns and curved glass. It appears within seconds of typing in a few keywords as if conjured from a coded mist. And, indeed, it was.
I am being guided through the world of artificial intelligence (AI) by Rob Armstrong, creative director of ThirtyC Yacht Design, who is showing me a few of the AI applications that his studio is currently using. Armstrong is one of the many yacht designers and naval architects who have been exploring AI as a tool to aid in the superyacht design process. Other designers have been less keen on embracing AI, warning it could ring the death knell for creativity.
The yachting industry’s varied opinions on AI are emblematic of the debate in wider society. There are those who are keen to dive in and see how it might be useful – and profitable – to humans. And there are those who fear its very existence, that its future iterations will spell an end to humankind as we know it.
Designers like Armstrong are experimenting with generative AI, which refers to algorithms that can create content such as images or text from amalgamations of information already available online. The much-discussed ChatGPT is one example. Artificial general intelligence (AGI), on the other hand, is the stuff of sci-fi: an advanced AI that, in the future, will be able to learn and think like a human and carry out a range of tasks, without human intervention. Some suggest generative AI is a stepping stone to AGI.
It is especially eerie when warnings on the dangers of advanced AI come from tech giants such as Elon Musk, who has said it has the potential for “civilisation destruction” and has called for governmental regulation (notably, China has announced its first regulations on generative AI). Musk, along with other tech leaders, including Apple co-founder Steve Wozniak and cognitive scientist Gary Marcus, signed an open letter in April 2023 calling for a pause in the “out of control” race for AI development, a warning that has yet to be heeded.
But AI is already in the mainstream and must be reckoned with, even by the yachting world. “People are afraid of AI because they see this robot who does his own thing,” says Marnix Hoekstra, co-creative director of Vripack. “But it’s a tool – like virtual reality or 3D printing – and it doesn’t do anything by itself.” The Dutch design house is currently exploring generative AI, such as ChatGPT, Midjourney and Firefly, a new version of Adobe Photoshop that has an AI plugin.
Winch Design is also experimenting with generative AI in its workflow and has had some interesting outcomes using software such as Midjourney, Stable Diffusion and Dall-E to generate reference images to kick-start ideas or conversations. “You can argue that inputting a written text prompt into Midjourney to generate an image isn’t that far removed from tasking a design team in the studio to create an image from a verbal briefing,” says Greig Jolly, partner, yachts and aviation, with Winch Design.
Back to that pyramid yacht. Armstrong is using Midjourney to create images by inputting text descriptions, such as “seashell-inspired-superyacht” and “Japanese-style-superyacht-interior”, and within moments images approximating these appear before us. Some are interesting, some are comical, but Armstrong says that, even with the latter, inspiration can be gleaned. “Designers see these images from a different perspective,” he says, referring to the pyramid yacht. “I’m looking at the colours – I love the turquoise and gold together – and how the superstructure has a soft blend on the top edge. I’m never going to use the whole thing, but I can use it for inspiration. AI is just a tool – software changes and you need to keep up with it.”
The seashell-inspired yacht prompt seems to blend a cowrie shell with a design from the late Zaha Hadid’s playbook. The “yacht” depicted here is not seaworthy, but Armstrong instantly points to the coral pattern on the aft deck ceiling as design fodder. He could use aspects of these images for ThirtyC client mood boards, which Armstrong says isn’t much different from clipping magazine pages or pinning on Pinterest.
In another application, Vizcom.ai, Armstrong inputs a line sketch he’s hand-drawn and it’s transformed into a fully formed yacht profile. He also mentions an open-source software called Blender that can take an AI-created image and translate it into 3D, which can be sculpted or drawn over. “But it’s a data-point 3D model, quite rough, not one with clean, symmetrical lines that you can actually bring to a shipyard,” he says.
Dasha Moranova Designs has employed AI in its Senses spa concept to play with general arrangements. “Senses’ sensory holistic design has been developed on Lateral Naval Architects' Innovative Free From Bulkheads (FFB) platform, which challenges traditional superyacht architecture,” says Moranova. The FFB platform enables watertight bulkheads to terminate at the lower deck, freeing up space and giving Moranova the opportunity to create an unusual open spa with almost 360-degree ocean views.
The interior is then built in a 3D model using Unreal Engine 5, MetaHumans and an AI conversational system. “This would be presented to owners wearing virtual reality [VR] goggles,” says Moranova. “The AI is so smart and sensitive that when you look at an element and focus on it for a few seconds, the AI will give any answer you might be thinking about.”
Vripack has regularly used virtual and augmented reality since 2014, and Hoekstra is excited by how AI will enhance these tools. “It will unlock a whole new level of customer experience – we can rapidly customise avatars and environments in VR, which will become massively lifelike and personalised and therefore more compelling,” he says.
In terms of visuals, both Armstrong and Jolly note that AI has been useful in quickly creating background images. “We can dream up an imaginary landscape or location and render it in AI almost instantly, then drop in the CGI model of our design,” says Jolly. “On a more granular level, the integration of AI functionality into Photoshop has relieved us of some minor, time-consuming image editing tasks.”
It’s easy to see how AI can be used in graphic design and animation. But a few of the top animators I spoke to didn’t want to comment, saying only that it’s too sensitive a topic for artists at the moment, raising concerns over copyright protection and job retention.
No copyright boundaries exist for the apparent plagiarism of original content on which generative AI subsists. These murky waters are full of ethical and moral questions, none of which will be solved swiftly. Sustainability is a concern as well, as generative AI, much like mining for cryptocurrency, relies on substantial computing power. “All the user sees is that an image pops up within seconds, but somewhere a huge computing farm probably swallowed up enough energy to power a small country,” says Christian Leyk, creative director at coquine![design].
Generative AI is, of course, derivative by nature, a fact that leads some designers, such as Leyk, to question its use in yacht design. “AI is not innovative, it’s not even design,” says Leyk. “It cannot create new concepts or ideas.”
Even Moranova, who has embraced AI in many ways, is wary of its use in creating images sourced from internet databases. “I don’t think reusing someone else’s property is the way forward for designers,” she says. “This is why true artists and designers have nothing to worry about. Emotional responses are a human trait and to provoke such a response you need to be human not only to interpret it but to evoke it.”
It’s the designers’ job, based on their real-world experience, to discern the viability of what AI spits out. “AI can create wonderful photorealistic images or professionally written text, but the user needs to have the knowledge to judge whether the information is correct and useful,” says Jim Robert Sluijter, lead exterior designer at Lürssen Yachts. “It’s similar with 3D modelling and photorealistic renderings, which can be a powerful tool to show a client what the yacht will look like. However, renderings can also show a very convincing and realistic image of something that will never work.”
Read More/Sustainably focused: the latest haute eco swaps in yacht designThe ideal scenario is that AI will shorten time spent on technical calculations so designers can spend more time on creativity. And this is the goal of Olesinski Ltd. The UK-based design and naval architecture firm is using AI in the most advanced way I encountered, having spent years and a huge investment to develop bespoke AI tools to optimise general arrangements and hull forms, by working with research teams at the University of Southampton. The studio has designs in build that have used AI which will be launched as early as 2024.
Bill Edwards, head of research and development at Olensinski, says the AI tool they’ve created is ideal for avoiding unnecessary trips around the design spiral. There are so many possibilities for creating a yacht layout or optimising a hull that designers can get quite far into the process before hitting a dead end and having to start over.
“The algorithm itself [of the Olesinski AI tool] is generic, there’s no explicit references to boat design, but it’s powerful at solving sorts of problems that require this type of computation,” says Edwards. “It can solve for several different objectives at the same time, whether furniture is placed correctly, cabins can be properly accessed, ceiling heights need adjusting… The software explores all the possible solutions, identifying dead ends, but does it rapidly, so flawed candidate designs need not be seen by human eyes. It gives us options.”
Previously, the Olesinski team would do a 2D CAD drawing, then fit it into a 3D model. “What we get out of AI now isn’t just 2D, it’s what we would’ve done with the 3D, which we can then manipulate, for example, if we want to move a bulkhead,” says managing director Justin Olesinski, who is quick to point out that what they produce with AI is only used during conceptual stages. “It’s not given to the yards to produce these lines. But at the level we need it to be, it’s as accurate as you’d ever want it to be.”
By feeding AI the results from simulations performed by potential hull forms, they can quickly create surrogate models and cast a wider net of ideas. “We can put in extreme features to investigate and are often surprised where we find performance improvements,” says Edwards. The AI might suggest a certain wave-piercing bow shape or suggest a chine width that the team didn’t necessarily expect.
“We will increasingly see AI doing several of the heavy lifting tasks in engineering design,” says Adam Sobey, associate professor in the maritime engineering group at the University of Southampton. “Creating a yacht layout takes a lot of time just to get to an initial concept. By using AI to develop initial concepts, humans can spend their time on the fine-tuning.” AI can help designers come up with half a dozen concepts early in the design process, allowing them to synthesise the best elements. AI can also be helpful in creating solutions to new regulations imposed by Maritime and Coastal Agency and other certification bodies.
AI is already transforming the study of design, but design fundamentals remain. “The AI approaches we have so far can’t pull together all the complexity of a real yacht,” says Professor Sobey. “It’s important that a student knows how to design a yacht first – the tools they use are less important. If they aren’t experts in traditional methods, how can they evaluate the designs that are being produced by the AI?”
Sluijter says that AI is no match for the years of experience that shipyards such as Lürssen possess. “The technical know-how on how to build, install and service the millions of components that go into a yacht, how they work together, and all the knowledge accumulated by building thousands of yachts by all the shipyards around the world over the past 150 years has never been shared on the internet,” he says. “No one will be able to build a yacht in the same way as a shipyard can by just using AI or the internet.”
Today, machine-learning modules are being included in maritime design studies, something that would traditionally be the realm for computer scientists. One of Professor Sobey’s students has gone on to work for Olesinski Ltd as a R&D engineer, bringing his knowledge of AI to a newly created role at the design house, showing that AI could lead to job creation.
“AI is a long way from stealing people’s jobs – we still require human designers for the detailed elements and to evaluate things such as aesthetics,” says Professor Sobey. “What we are doing is augmenting designers [and allowing them] to explore more of the design space than they were previously able to. This is both more and less exciting than is in the news. It certainly isn’t AI taking over the workplace, but it is providing new tools that allow us to interpret what we have been doing for many decades in totally new ways.”
But Leyk is less optimistic. “It’s very likely some designers will lose their jobs,” he believes. And Hoekstra notes that the World Economic Forum Future of Jobs Report 2023 predicted that nearly 25 per cent of jobs will be disrupted.
“Imagine that around a quarter of all the people you know will have their job disrupted over the next five years – that’s massive,” says Hoekstra. “So, obviously, there are ethical considerations for us.”
Hoekstra thinks that AI will inevitably disrupt yacht design as we know it. “And I honestly mean ‘disrupt’ because I have a strong belief that there will be two kinds of design studios at the end of this decade: those that are fully utilising AI and those that are out of business,” he says. “We either have to adopt it or die.”
Olesinski believes that the only downside to AI is the amount of time and capital a company needs to invest. “We’ve invested a lot over 10 to 13 years in hull forms and four to five years on the general arrangement side of things,” he says. “If a design house wants to do that, they need to have an R&D department specifically for AI.” Olesinski Ltd will be offering its AI tool as a paid service to other designers who haven’t made such an investment. A potential problem arises in clients thinking they should be charged less if the design process is seemingly quicker. But this overlooks the expertise and R&D that has gone into the back end; time that will be freed up to spend on creativity and perfecting a design.
Jolly posits that even owners themselves might have a play at AI. “The day will come when a client presents us with an image of their dream that they’ve made in AI and asks us to create it for them,” he says. “But AI won’t replicate the unique individual character and artistry that a craftsman brings to their work. I think that human touch and sense of tradition is something that is very much valued in our industry.”
Perhaps the answer to the AI question in yacht design is not a matter of all or nothing, or kill or be killed, but a strange hybrid, putting us one step closer to the singularity between man and machine that futurist Ray Kurzweil predicted nearly 20 years ago. Yet for now, yachting remains very much a human pleasure and a product of human creativity.