As an industrial designer, art director, and curator of educational programs, I can’t help but notice how the profession is changing. Over the past two years, neural networks have changed it faster than the entire previous decade. This article is an attempt to figure out what exactly is happening: with education, with the market, with those who are just entering the profession, and with those who are already working in it.
Easy entry into the profession
A new way to enter industrial design has appeared — fast and pleasant. You don’t have to learn sketching, spend hundreds of hours mastering 3D modeling and visualization, or understand the logic of building a form. It is enough to open a neural network for industrial designers, for example, Vizcom, feed it a quick sketch, write a prompt, and quickly get a spectacular result – visualizations that look quite convincing. And more recently, a polygonal 3D model.
These tools are simple to master, which lowers the barrier to entry into design. For the market, this means one important thing: the visual result has become cheaper and faster. Those who used to order a cheaper design — a freelancer from an exchange or a novice student — can now get by with a neural network.
Previously, there was a path behind any such picture — through analysis, sketching, modeling, visualization. Through dozens of bad decisions. Through constant collision with limitations. This path formed an understanding and directly influenced the quality of future products.
Those designers who have already completed it will continue to work – the difficulties of mastering skills are already behind. But students who are just mastering the profession find themselves in a different situation. I already hear from my colleagues that their students complete projects almost entirely through neural networks — without building 3D models or making visualizations manually. If the stage of work through sketching and modeling falls out of the design process, the expertise that these skills form may disappear along with it.
Neural networks are deprived of skills
This phenomenon already has a name and a research base. When you look for ideas in sketches, build a form, make mistakes and correct mistakes, you gradually acquire a unique style, aesthetic compass, mastery of composition, attention to detail. Thanks to neural networks, you can get a result bypassing the path to it. Because of this, the most difficult skills of a designer to develop are at risk of going out of circulation.
This effect has already been called deskilling. Unused skills fade away. Studies have shown that doctors’ diagnostic skills decreased after just a few months of working with AI — and they themselves did not notice it.
Even more alarming is a 2025 study that coined the term “creative scar”: while generative AI can boost creativity, users quickly lose it as soon as they stop using it. It is becoming lower than before.
In design, these processes can remain invisible for a long time — until the moment when the graduate first encounters a task in which the neural network will not be as skilled as in simple tasks and products. This moment has not yet come, but we may face it in a few years.
A new type of designer
On this basis, a new type of specialists may emerge — those who can quickly generate concepts, produce aesthetic images, and work with AI as the main tool. The question is whether the market needs such a specialist.
Judging by the real vacancies, not yet. Employers still require CAD, 3D modeling, visualization, understanding of manufacturing techniques. Vizcom and other AI tools, if they appear in vacancies, are in addition to these skills, not instead of them. A beautiful AI portfolio is not an advantage — high visual quality has always been the industry standard. And in the new conditions, photos of layouts, screenshots of CAD models — everything that shows the level of skills without AI — can be a plus.
Professor of Industrial Design at NC State University Byungsoo Kim, studying students with AI tools, found that AI generates average results and has difficulty creating original solutions. At the same time, students are divided into two camps: conservatives, who are cautious with AI and strive for mastery, and enthusiasts, who use AI in all possible tasks.
Although other niches may open up for some of these specialists — not in industrial design, but in media, content, and fashion. Influencer designers, visual futurists, authors of conceptual provocations are those who generate images quickly and convincingly. But this is a different profession. The question is whether the students themselves understand this – and whether the education system that trains them understands it.
Education at a crossroads
I am in charge of educational programs, so this question is not abstract for me. Universities are actively integrating work with neural networks into the educational process — this is a logical result of the desire to be relevant. But here a trap arises: students begin to use neural networks not only where it is provided, but also everywhere where it gives results faster. And the system does not yet have time to react to this — and can highly appreciate works that look convincing thanks to AI, but do not reflect the real development of the student.
I’ve already seen projects at university exhibitions without a 3D model, mock-up, or prototype — only with crude neurovisualizations. If the work done independently and the work generated by the neural network receive the same grade, the motivation to follow a difficult path disappears.
The responsibility of teachers is to evaluate not only the result, but also the process, to fix specific expectations for the use of neural networks in the criteria. Otherwise, the system will produce specialists who are not ready for the reality with which they will have to work.
Design without designers
The same logic works at the business level. An engineer, designer, or product manager opens Vizcom, gets options for the appearance of the product in a short time — and from a business point of view, everything looks reasonable: faster, cheaper, no need to involve anyone. In November 2025, Vizcom raised $27 million in investments with the ambition to become the “Figma for the physical world” – from the first sketch to production, including 3D modeling, material simulation and production verification. The tool becomes more powerful.
But even the images obtained by neural networks need to be worked out for the concepts to be applicable. And after the concept stage — when images need to be turned into a detailed 3D model, work out every surface, every transition, colors, textures, product graphics — neural networks are much less useful.
In the process of design, the design changes to meet technical realities. And here we need designers who accompany these changes and make sure that the product does not lose itself on the way from the screen to production. McKinsey confirms that AI speeds up the process, but it does not replace professionals. Without an expert behind the wheel, tools can even harm the project.
What will happen to the market
All this affects the market — and not as unambiguously as it seems. Neural networks do not make the entire design cheaper. They make cheaper design even cheaper. Companies that used to save on design can now save even more — it’s just that the tool has changed.
Some industries are already rebuilding processes: AI visualizations developed by industrial designers go directly into production, bypassing the stage of a full-fledged 3D model. For some product categories, this works. Projects are accelerating, their circulation is growing.
But at the other end of the market, the opposite can happen. When the visual result has become cheap, the value of what neural networks cannot do may increase. The ability to distinguish a working solution from a beautiful one. Understand where form argues with function. Ensure that the product is fit for purpose from screen to production.
These are still forecasts, not facts. But the history of similar shifts provides several scenarios at once, and all are plausible. The overall design budget of the market may be reduced. The market will most likely not polarize, but will be restructured: some of the tasks will go to tools, new services will appear, for example, “AI design concept” from professional studios or “design support for your AI concept”.
Many companies will go where it is faster and cheaper. The question is what will remain at the other end.
How to proceed
Neural networks will not disappear. The question is not whether to use them or not, but how to use them.
If you are a student, the main task is not to deceive yourself. Neural networks can hide gaps in sketching, modeling, or visualization, speed up work, but they will not form thinking. The ability to make a good design comes “through calluses”. You should not stop working with your hands: draw, build 3D, try to assemble the shape yourself. It is in these attempts, often unsuccessful, that the expertise grows, which ultimately allows you to obtain an outstanding result. Sketching is not archaic here: it is an important tool through which a sense of form, proportions, and composition is formed. The hand that draws thinks. Don’t let neural networks do the math for you.
If you are a practicing designer, use neural networks, if you haven’t started yet, it’s high time. Experiment and find ways to use them creatively. Let neural networks become nothing more than a tool that speeds up the process and allows you to do what was previously impossible. Personally, I see the most benefit not in Vizcom at all, but in the AI assistant — an endless source of knowledge and expertise. I turn to him every day.
If you are an entrepreneur or an engineer, play with neural networks, try to go through the path of creating a product without a designer. Most likely, you will see for yourself where the tool ends and the work that it does not know how to do begins. Even better: hire a good designer and pay for a subscription to the neural network that he advises. Well, if what a neural network without a designer will give you is enough, great, thanks to neural networks, they saved everyone time and nerves.
The tools have always changed. The speed changed, the availability changed. One thing did not change: behind any result there is a person who understands what he is doing – or does not understand. And this is always visible.
If you are developing a product and looking for a design studio, this is what we do, and you can follow our work and new materials on Telegram.
