A study of GenAI usage by Design Students Analysis of Survey Results and Journals of AI practices at the Politecnico di Milano in 2025/2026

Imagine the world of design as a massive, bustling construction site. Before a single brick is laid or a beam is raised, the architects must spend hours digging through mountains of blueprints, studying the soil, and talking to the people who will live in the building. This "digging" is called research, and it's the foundation of any great creation. Now, imagine a new kind of super-fast, super-smart robot assistant has just arrived on the site. This robot, powered by Generative AI (or GenAI for short), can read thousands of books in a second, summarize complex reports, and even sketch rough ideas. The big question everyone is asking isn't just "Can the robot do the work?" but "Does using the robot make the architect feel like they aren't the real creator anymore?" And more importantly, does the robot help the architect build a better house, or does it just make the process faster without teaching them anything? This is the exact mystery a team of researchers at the Politecnico di Milano set out to solve by watching how design students in Italy are using these digital assistants.

The researchers, led by Stefana Broadbent, decided to peek behind the curtain of the 2025/2026 school year to see what was really happening. They didn't just ask students if they used AI; they asked how they used it, when they used it, and whether it made them feel like their own ideas were being stolen by the machine. They surveyed 280 design students and asked a specific class of 100 masters students to keep a detailed diary of their AI usage while working on a project about water scarcity in Sicily.

Here is what they found, and it's a bit of a surprise.

The Robot is a Research Assistant, Not a Magic Wand
The students are using GenAI like a super-powered librarian and editor, but they are mostly keeping the "magic" for themselves. The data shows that students use these tools heavily during the early stages of a project: finding information, summarizing long articles, brainstorming ideas, and fixing grammar. It's like having a friend who can read a 500-page book in five minutes and tell you the main points, or who can instantly translate a foreign language document.

However, when it comes to the actual "making" of the design—like generating images, writing code, or building prototypes—the students are surprisingly cautious. In fact, less than one-third of the students reported using AI to generate images or music. They aren't letting the robot paint the picture; they are using it to find the reference photos and then painting it themselves.

The "Trust but Verify" Rule
One of the most interesting discoveries is that the students don't blindly trust the robot. In fact, they are quite skeptical. About 65% of the students said they don't fully trust the AI's answers, and a whopping 85% said they systematically edit and change what the AI produces before using it.

Think of it like this: The AI is a very fast, very chatty intern who is eager to please but sometimes makes things up or gets details wrong. The students treat the AI like a first draft generator. They ask the intern, "Hey, give me a list of sources," and the intern hands over a stack of papers. But the student immediately checks every single paper to make sure it's real and relevant. They don't just copy-paste the intern's work; they use it as a starting point to build something better. This constant checking and editing means the students feel they are still the captains of the ship, steering the direction even if the AI is helping with the rowing.

Creativity and Ownership: Still Safe
Perhaps the biggest worry for teachers and parents is that using AI will make students feel less creative or like they don't own their projects. The study suggests this fear might be overblown, at least for now. Only 31% of the students felt that using AI made them feel less creative or that they had lost ownership of their work.

The students seem to have found a sweet spot. They use AI to speed up the boring, time-consuming parts of research—like finding articles or transcribing interviews—so they can spend more time on the parts that require human judgment, like interpreting what those interviews mean or deciding how to solve the problem. Because they are doing the heavy lifting of thinking and deciding, they still feel like the true authors of their projects. The AI is a tool that makes things faster and easier, but it isn't replacing the student's brain.

The "Sicily Water" Diary
To get a closer look, the researchers asked a group of students working on a project about water scarcity in Sicily to keep a journal. These diaries confirmed the survey results. The students used AI to brainstorm research questions, find hard-to-locate government reports, and summarize long academic papers. They even used it to transcribe interviews with local farmers. But when it came to the actual analysis—figuring out why the water was scarce and what to do about it—the students took over. They didn't ask the AI to solve the problem; they asked it to help them gather the clues.

What Does This Mean for the Future?
The study suggests that design students are becoming "reflective experimenters." They aren't just using AI because it's cool; they are figuring out exactly where it fits in their creative process. They use it to clear the path so they can run faster, but they are still the ones running the race.

However, the researchers warn that this balance is delicate. If students get too comfortable with the AI doing the research, they might stop learning how to dig deep into the information themselves. The study raises a big question for teachers: How do we make sure students keep valuing the hard work of research, even when a robot can do it in a flash? The answer might be to teach students not just how to use the tools, but how to build their own custom tools, turning them from passive consumers of AI into active creators of their own digital assistants.

In short, the design students of 2025/2026 aren't letting the AI take the wheel. They are using it as a powerful co-pilot, checking its map, correcting its course, and making sure they are still the ones deciding where the journey goes.