AI and Creativity: Why Yes, Why Not

A few minutes of scrolling is all it takes to notice. Renders bathed in the same golden light, sand-toned interiors with the obligatory plant in the corner, captions that all open the same way. The names of the studios change; the point of view rarely does. By 2026, artificial intelligence has become a fixture of designers’ daily work, and social media was the first place its effects became visible: images, texts and colour palettes are converging on a shared aesthetic, as if the entire sector had entrusted its identity to the same art director.

Yet it would be a mistake to dismiss the matter as a rejection of the tool. Used with judgement, AI for architects is today one of the most valuable resources available to an architecture studio. The real question, then, is not whether to adopt it, but where to place it in the work process. And, above all, at which moments not to leave it on its own.

Why Yes: the invisible, unpaid work of an architecture studio

Every architecture studio knows the gap between the work that creates value and everything that surrounds it. On one side are thinking, designing, deciding. On the other, a steady stream of tasks that are necessary but unproductive: the schedule to realign for the third time in a single week, supplier quotes to chase, the costs of a building site to reconstruct after the fact, the right drawing to dig out from among hundreds of files. No client pays for them, no project gets better because of them. Yet they absorb precious hours.

This is precisely the ground on which artificial intelligence shows its potential.

1. The planner

Once the key dates of a commission are set, it can build the project timeline in a few minutes, with phases, deliverables and site milestones, and keep it up to date within a space shared with collaborators and consultants, without every change forcing you to start over.

2. The studio controller

A kind of constantly updated business plan that compares site expenses against the budget, tracks fixed costs and the hours spent on each commission, down to personal expenses, flagging early where something risks slipping away.

3. The archivist

On the most complex projects it proves a reliable ally. Drawings, renders, site photographs, technical sheets and minutes can be organised in a shared folder according to a clear logic, accessible to the whole team.

4. The synthesiser

When a forty-page competition brief arrives, or a building code, or the minutes of a long meeting, AI returns a reasoned summary, leaving the designer time to focus on decisions.

5. The assistant who never sleeps

It handles day-to-day correspondence, from requests for quotes to reminders, all the way to messages sent to dozens of contacts at once. An assistant who keeps no office hours, whose drafts still need to be read and signed.

6. The sparring partner

Perhaps the least obvious role: a critical interlocutor to think out loud with about a concept, to put the arguments of a presentation in order, to anticipate the objections a jury might raise.

7. The proofreader and format translator

Project reports, award submissions, website copy, translations: AI corrects typos and inconsistencies and can turn a technical report into a short project sheet, a caption, a post. But even when a studio chooses to handle its own communication in-house, it needs a dedicated person who brings ideas and vision. The tool can refine the form; the content has to come from somewhere else.

In all these cases the same principle applies: the material already exists, and it belongs to the studio. Artificial intelligence organises it, checks it, makes it more accessible. And it gives back time, which can return to where it is truly needed: the project.

Why Not: when the tool becomes the author

The skipped step

The problem begins when a step is skipped. Artificial intelligence is no longer asked to help express an idea, but to supply the idea itself. A post about the new project, a mood image for the house by the sea, a studio presentation text: generic requests that produce formally flawless results, often welcomed with enthusiasm by those who commissioned them.

The drift towards the average

Seen from the outside, however, the same content tells a different story. Anyone who comes across dozens of these pieces every day immediately recognises:

• the same adjectives and the same sentence rhythms;
• the same low sunset light;
• the same palette of sand and terracotta.

This is no coincidence. A language or image model, without precise instructions, naturally tends towards the most probable solution, that is, the average. And the average, by definition, is the same for everyone.

The loss of trust

For the reader, the effect is first inattention, then something deeper. You stop reading to the end, you stop following, and eventually you stop trusting. Because a text that anyone could have written leaves the suspicion that there is no one behind it. And that is a suspicion an architecture studio cannot afford: what it offers, ultimately, is precisely a point of view, a method, a person.

The Difference: treating it like a collaborator

Artificial intelligence is truly useful only to those who already know their craft. The most fitting comparison is with a new collaborator who has just joined the studio: well trained, extremely fast, but completely unaware of who we are and what we are looking for. Ask them simply to make something beautiful, and they will deliver the most generic solution they know. Explain precisely what is needed, why, for whom and within which constraints, and the result changes radically.

1. Start from the studio’s own idea

Before opening any tool, it is worth setting down in a few lines what you want to communicate and why it matters. Without that clarity upstream, no algorithm can make up for it.

2. Provide authentic material

Site notes, the project report, something the client said, your own photographs. The more specific the starting point, the less the result will slide towards the average.

3. Give instructions like an art director

Not like a casual user: the tone, the length, the words to avoid, the references to look at.

4. Use images with caution

To explore a direction or build an internal moodboard, image generation can be a legitimate tool. To tell the story of a built work, however, real photographs remain irreplaceable, because they are the only proof that the project exists.

5. Take back the final step

The step that comes before publication must always return to whoever signs the work. The draft may be born elsewhere; the final version must carry the studio’s own words and rhythm.

The principle, in the end, is simple. Artificial intelligence executes; the architect decides. When the roles are reversed, the work becomes faster, but it stops being truly theirs.

Thinking stays human

No architect would leave it to modelling software to decide the shape of a building. It is used to draw, faster and more precisely, what has already been imagined. The same principle applies to artificial intelligence: an extraordinary tool for freeing up time from emails, checks, planning, summaries and organisation, but not a substitute for what makes a studio recognisable: its view of the project, the way it chooses to tell it, the images it uses to show it.

In a landscape where everything tends to look alike, the true rarity becomes an authentic voice. That is what trust is built on. And it is the one thing no tool will ever be able to generate in our place.

Want to understand how to bring AI into your studio’s communication without losing your voice? Get in touch: thus-newswire.com