Fresh ideas from museums around the globe in your inbox each week

In its Warm-up with AI workshops, Kunsthalle Praha’s learning team has students use tools like ChatGPT and Runway to respond to works in its exhibitions, writing stories, animating photographs and then comparing what the AI made with the original art.
AI is often seen as the enemy of creativity — a machine poised to replace the artist, the writer, the dreamer. But what if it isn’t replacing creativity at all? What if it’s becoming its newest catalyst?
That’s the question being explored inside an unlikely setting: a former 1930s transformer station in the centre of Prague. Reimagined as Kunsthalle Praha, the space hums with the same energy it once distributed across the city — only now, it’s fuelling experiments at the intersection of art, technology, and learning.
For the museum’s Learning and Public Programmes team, led by Martina Freitag and Barbora Škaloudová, AI is a creative material — something to stretch, bend, and test. And their workshop series, Warm-up with AI, invites students to do exactly that.
Each reported session lasts two hours: 30 minutes of artistic context, an hour of group creation and a final half-hour of reflection.

“Think of it like stretching before exercise,” says Freitag. “You don’t start by running a marathon. You start by testing your limits.”
The workshops begin with simple, playful interactions with AI. Students use image and text generation tools to create short pieces inspired by the museum’s exhibitions. But beneath the fun lies a deeper aim: to shift perception from fear to familiarity.
“When photography first appeared,” Freitag explains, “many artists were afraid of it. But once they learned to use it, entirely new art movements emerged. AI feels the same right now.”
In other words, Kunsthalle Praha is teaching young people not just how to use AI — but how to think with it.

The idea first took shape during the 2023 exhibition READ by artists Elmgreen and Dragset. Known for their conceptual installations, the duo transformed a gallery into a library — except the books were turned backwards, their titles hidden.
Visitors were puzzled. What were they supposed to read when there was nothing to read?
Freitag turned that confusion into a prompt. Literally.
Using ChatGPT, participants wrote their own literary interpretations of the exhibition. They selected an artwork, chose a genre, and asked the AI to generate a story around it. The results were as diverse as the imaginations behind them — detective tales, romantic monologues, surreal fables. Each was a small act of co-creation between human and machine, the algorithm acting less as author and more as accomplice.
The educators describe the exercise as a way to turn observation into active interpretation, with participants making choices about how to respond to the artwork.

The next exhibition couldn’t have been more different. Gone were the conceptual installations; in their place came the crisp, disciplined photography of Lucia Moholy, a lesser-known figure of the Bauhaus.
Her story is both extraordinary and unjust. Moholy’s negatives were taken by Walter Gropius when he fled Nazi Germany, leaving her without access to the very work that defined her career. For decades, her images circulated without credit, her role in the modernist movement almost erased.
Freitag and Škaloudová saw a parallel between those lost negatives and the blind spots of history — and they wondered: could AI help students reimagine what’s missing?
In the reported workshop, they used Runway, which can generate moving images from still photographs. Students selected one of Moholy’s images and used AI to create a short video imagining what might have happened just before or after the shot.
These were speculative animations inspired by the photographs, not reconstructions of what historically happened before or after the shutter opened.
“AI often made the portraits too perfect,” Škaloudová recalls. “It erased the individuality, smoothing away the human details. So we guided students to think differently — to use AI to suggest a moment, not rewrite it.”
The most powerful outcomes, she says, were the subtle ones: a shadow shifting, a gesture continuing, a line of architecture breathing to life.
The same distinction matters when considering how museums present AI-generated imagery.

Each Warm-up with AI session lasts two hours. The first half-hour introduces artists working with technology, the middle hour is hands-on creation, and the final 30 minutes is reflection — not critique in the traditional sense, but a shared conversation about process and perception.
Participants work in groups, like small film crews. One student might compose the shot, another write the prompt, a third oversee the editing. The process demands clarity of thought — how do you describe an idea precisely enough for a machine to understand?
That discipline, the educators say, is what students carry away most strongly. They learn to translate vision into language, to think about what a camera means to do, not just what it shows.
“Writing a good AI prompt,” Škaloudová notes, “isn’t so different from directing a film. You have to know what story you’re trying to tell.”

As Kunsthalle Praha’s exhibitions change, the workshops evolve alongside them. When exploring Chiharu Shiota’s installations of thread and shadow, students focused on movement within space — on how a single camera angle could suggest a story without disrupting the fragile stillness of the work.
At another exhibition, Bergman and Hartung and We’ll Never Be Parted, participants worked with abstract paintings, using AI to explore the gestures and textures that imply motion. The challenge, Freitag says, was maintaining the original artist’s spirit.
That attention to the source connects with the V&A’s approach to creative fidelity.
“The AI sometimes made things too literal,” she explains. “A swirling brushstroke became a spinning vortex. We had to remind students that abstraction isn’t meant to explain — it’s meant to evoke.”
In other words, AI wasn’t a shortcut to meaning; it was a mirror for interpretation.

The final half-hour matters because the generated image is only one part of the exercise. In the Moholy workshop, an unexpectedly polished face prompted questions about the individuality the original photograph held. With abstract painting, a movement made too literal created a different problem: the tool had narrowed an open-ended work. Returning to the source lets students describe those changes and explain their choices. The workshop’s sequence therefore keeps looking, making and discussion connected, even when the software produces something visually convincing.
What emerges from these sessions isn’t just digital output, but a new way of thinking about learning itself. Freitag and Škaloudová act less like teachers and more like moderators, guiding conversations rather than dictating lessons.
“There is not one all-knowing guide,” says Freitag. “Learning is a two-way street.”
The approach has turned Kunsthalle Praha into a kind of laboratory — one where art and technology continually test each other’s limits. And in that process, both educators and students rediscover a crucial truth: innovation is rarely about the tools. It’s about the questions you ask with them.

The workshops keep returning to the questions students ask of the artworks and the choices they make with the tools.
Because when students stop fearing the machine and start playing with it, something remarkable happens. They begin to see art — and the future — through new eyes.
–
This article is based on a presentation by Martina Freitag and Barbora Škaloudová from Kunsthalle Praha given at the Creative Museums Summit on October 22 2025.
Manuel Charr is a journalist working in the arts and cultural sectors. With a background in marketing, Manuel is drawn to arts organizations which are prepared to try inventive ways to reach new audiences.
Fresh ideas from museums around the globe in your inbox each week