Exploring generative art with Francesco D’Isa
Creativity is a fundamental human trait, and what evolves is the way we engage with it.
This is the first AI Muse interview, and I couldn’t think of a better person to start with Francesco D’Isa for several reasons. First — full disclosure — we’ve connected through shared spaces on social media, where we’ve often found ourselves aligned in expressing doubts about the rigid, often vocal opposition from parts of the art world to AI technologies.
We both recognize the real challenges and ethical concerns that come with AI, but at the same time, we believe in embracing these tools to their fullest potential—for everyone. This balance of skepticism and openness makes Francesco an ideal first guest to explore the intersection of creativity, philosophy, and generative AI.
Francesco D’Isa, born in Florence in 1980, is an Italian philosopher, writer, and digital artist whose work bridges traditional and cutting-edge creative practices. He has exhibited internationally and published graphic novels, essays, and novels with prominent publishers. His latest works include L’assurda evidenza (2022) and Sunyata (2023), a graphic novel featuring AI-generated images created with tools like Midjourney and Stable Diffusion.
This innovative approach earned him both acclaim and controversy, with some critics accusing him of “stealing images” and bombarding him with insults and threats —questions he tackles with his usual philosophical clarity and wit.
Let’s start with Francesco our journey across generative art: this will bring us meeting a lot of interesting artists and their creations.
The role of AI in art
How do you think AI is changing the concepts of creativity and authorship in the art world?
I don’t believe that creativity itself changes with the advent of new tools. Creativity is a fundamental human trait, and what evolves is the way we engage with it. Each new tool opens up pathways that were previously unimaginable, but creativity adapts—creatively, of course — to these new means and solutions.
AI, in particular, reveals the deeply collaborative nature of all forms of creativity and authorship. It serves as yet another blow to the Romantic ideal of the isolated genius—a concept that, despite its limited historical and geographical origins in Romanticism, continues to exert a disproportionate influence on contemporary thought.
The idea of the lone, inspired genius emerged primarily in 19th-century Europe, where it became a cultural cornerstone. Yet, the 20th century, particularly through contemporary art and postmodernism, mounted significant critiques of the notion of the author. Figures like Roland Barthes and Michel Foucault paved the way for understanding creativity as a collective and contextual phenomenon. AI takes this critique further by materializing it — literally showing how creation involves networks, datasets, and interactions.
In short, AI doesn’t change creativity itself but illuminates its true nature: adaptive, collaborative, and constantly evolving. It offers a mirror to our artistic processes, reminding us that no creation exists in isolation.
Technical and creative aspects
How do you conceptualize a new piece of AI-generated art? Could you walk us through your process, from idea to execution?
My approach to creating AI-generated art is deeply rooted in co-creation with the tool itself. I see AI not as a subordinate assistant but as an active collaborator. The process begins with exploration rather than a fixed plan—an open-ended dialogue with the technology.A key aspect of my method is embracing errors. Often, mistakes or unexpected outputs reveal alternative pathways that I hadn’t considered, leading to entirely new directions for the work. This openness aligns with a somewhat anarchic creative philosophy: the plan for the piece emerges gradually, rather than being rigidly defined beforehand.
This way of working isn’t unique to my AI-based projects — it’s also how I approach creation in general, regardless of the medium. AI amplifies this dynamic by offering an unpredictability that mirrors human creativity in fascinating ways. The algorithm’s “errors” become opportunities, and the iterative process of refining prompts, inputs, and outputs becomes an integral part of the creative journey.Can you share a specific project where AI significantly enhanced your artistic vision? What tools or algorithms did you use?
My graphic novel Sunyata (Eris edizioni) stands out as a project where AI profoundly shaped my artistic vision. The story is deeply personal, rooted in my anticipation of a profound loss and the way I navigated it through (also) the philosophical tools of Zen Buddhism. What made the creation of Sunyata so distinctive was the parallel development of the narrative and the imagery—each influenced and shaped the other, creating a dialogue between words and visuals.
In terms of the imagery, I began by setting a specific stylistic foundation. Once this was established, my involvement in directing the AI became deliberately minimal. I liken the process to drawing tarot cards and interpreting their meanings: I would provide a prompt or input, allow the AI to generate visual elements, and then reflect on the results to extract meaning and direction. This method felt almost meditative, allowing the unexpected outputs to inspire narrative shifts or deepen the themes of the story.
The tools I used were primarily generative TTI models designed for visual creation like Midjourney or Stable Diffusion. The result is a work that feels deeply collaborative—a fusion of human intuition and algorithmic unpredictability. It mirrors the Zen perspective of embracing the present moment and finding beauty in the unplanned.
How do you ensure that the use of AI in your work remains innovative and doesn’t become repetitive or predictable?
The truth is, I can’t fully ensure that it doesn’t! This is especially true in a context where AI technologies are developed by third parties. Using AI in creative work comes with an inherent dilemma.
On one hand, open-source AI tools offer much more freedom and customization. They allow for a deeper exploration of the technology’s potential, but personalizing these tools is often incredibly complex—I’d even say impossible beyond a certain point. The technical demands can be so high that they risk stifling the creative process entirely.
On the other hand, commercial AI tools are far easier to use, requiring neither advanced hardware nor extensive programming skills. However, these tools come with a price and limitations imposed by the companies that develop them. They typically offer minimal options for modifying the underlying software, which can constrain creativity and make the outputs feel more standardized.
There are some promising middle-ground solutions. Tools like MidJourney, for instance, strike a balance by being relatively user-friendly while still offering a degree of flexibility and innovation. However, we are still in the early days of generative AI tools, and there isn’t yet a true “Photoshop AI”—a platform that is intuitive to learn but also provides extensive functionality and creative freedom.
What role does intuition play in your work with AI, and how do you balance it with the algorithmic nature of the technology?
To answer this, I first need to define intuition. If we think of intuition as a kind of unconscious, out-of-the-box thinking, then I would argue that intuition isn’t exclusive to humans—algorithms can exhibit a form of intuition as well. It often manifests in the form of errors or unexpected outputs.
Interestingly, these “errors” are not so rare. Every mistake an algorithm makes could be seen as a kind of intuition—a deviation from the expected pattern. However, in commercial AI, these errors are minimized to make the software more predictable and user-friendly for the general public. The result is a system that works smoothly but sometimes at the cost of creativity.
For both human and algorithmic intuition, there are two essential elements: making mistakes and recognizing which of those mistakes might actually be solutions. In my work, I embrace these errors — both my own and the AI’s — as new possibilities. My role is to spot when something unexpected carries potential and then develop it further.
Ethical and philosophical considerations
What ethical aspects do you take into account when creating AI-driven art?
Answering this question thoroughly would require a book — and, in fact, I’ve written one, though it addresses broader themes. The ethical issues I consider when working with AI differ significantly from those typically discussed by artists in this field. For instance, intellectual property is not a concern for me at all.
Every artwork is inherently a collective work. Intellectual property, in my view, is merely an economic device — and a flawed one at that. It tends to benefit corporations at the expense of individual artists. The anxiety around AI “stealing” from human creators seems misplaced to me, as it overlooks the fundamentally collaborative and derivative nature of all creative processes. Art has always been built on shared knowledge, techniques, and influences.
The ethical issues I find more urgent are those related to environmental impact and the monopolization or restriction of AI tools. AI technologies, particularly those based on large-scale machine learning, require immense computational resources, which contribute to carbon emissions and environmental degradation. This is a pressing concern, especially as the use of these tools becomes more widespread. Equally troubling is the concentration of AI development in the hands of a few corporations. These companies dictate how the technology can be built and used and who has access to it, creating barriers for independent creators.
These issues, however, are not exclusive to AI — they reflect broader systemic challenges in the technological and economic landscape. AI reveals old problems more than it creates new ones.
What are your thoughts on the exploitation of human creativity by major AI companies?
I don’t believe this is happening — at least not in the way people often fear. The exploitation of human creativity by large companies occurred long before the advent of AI. These corporations have been leveraging creative work for decades, and the primary mechanism for doing so has been copyright — the very system many creatives now invoke in an attempt to defend the small crumbs they were previously allowed to keep. Copyright, in its current form, doesn’t protect artists. Instead, it serves as a tool for corporations to consolidate control over creative content, limiting its circulation and profiting disproportionately from it. AI tools have simply exposed the flaws of this system more clearly.
The real ethical issue with major AI companies lies elsewhere. These corporations offer closed systems, restricting the ways in which AI can be used and accessed. They aim not just to dominate the market for these tools but also to control the ethical governance of their use. By defining both the product and the rules of engagement, they position themselves as gatekeepers, dictating how and by whom AI can be utilized.
This concentration of power is far more troubling than concerns about creativity being “exploited.” It risks stifling diversity in AI applications while granting a few corporations unprecedented influence over cultural production.
The future of generative art
How do you see the role of AI evolving in the art world over the next decade?
If I knew the answer to that, I’d probably be a better artist than I am! That said, I think AI will likely become just another tool among many, much like photography, painting, or computer graphics. It will find its place as one of the many techniques artists can choose to work with, depending on their personal style and vision. Not everyone will use AI in their practice, and that’s perfectly fine — but everyone will have to engage with it in some way.
What challenges and opportunities do you foresee for artists working with AI in the future?
One of the biggest challenges will be ensuring that AI tools are both accessible to everyone and free from the constraints of corporate control. Currently, many AI systems are either proprietary and commercial — designed to limit how the technology can be used — or highly technical, requiring significant expertise and resources to customize and operate.
This creates a dual form of gatekeeping. On one side, corporations impose restrictions on widely available tools, limiting creative freedom. On the other side, artists who can afford custom AI setups and possess the technical knowledge to use them often become part of an exclusivity. The complexity and cost of these personalized systems create barriers that leave many creators behind.
The ideal solution lies in finding a balance. AI tools need to be accessible and open enough to allow wide participation but without becoming so commercialized or oversimplified that they sacrifice creative flexibility.
Personal insights
What has been your most rewarding experience as an AI artist, and why?
The most rewarding experience has been getting to know new creative people and, perhaps even more importantly, discovering new aspects of myself and my artistic practice. Working with AI forced me to analyze and question my approach, much like the way we confront ourselves during moments of upheaval or “trauma” — in this case, the disruption brought by a powerful new tool.
How has working with AI changed your perspective on traditional art forms?
To be honest, it hasn’t changed my perspective on traditional art forms at all. I don’t see AI and traditional art as being in competition or mutually exclusive. What will change, however, is the future of traditional art. When polished, commercial-grade images no longer require significant resources but only minimal effort, the value and perception of such imagery will inevitably shift. This hasn’t fully happened yet, but it will. The evolution of AI will reshape how we view and define the role of traditional techniques in a world where technical perfection is so easily attainable.
Last Shots
Can you name an AI artist we should interview?
Eugenio Marongiu, also known as Katsuko Koiso, is an excellent choice. A Milan-based artist, he merges AI with photography to explore innovative storytelling, creating visually striking works and videos.
Did you use an AI assistant to answer this interview?
Absolutely, yes. I provided the answers in Italian, quickly and in a rough form, and the AI refined them into polished responses in English. When something didn’t align with my thoughts, I adjusted it. And I’m certain you used AI to craft these questions, didn’t you?
You got me! But it would have been silly not to.
🛠️ Top AI Tools for Video Avatar Creation
HeyGen offers a versatile platform for generating AI avatars with features like: Instant avatar creation for quick videos; sudio-quality avatars in 4K resolution; multi-language support with over 40 languages; integration with streaming software for live content.
Synthesia is a leading platform for creating lifelike AI avatars, particularly suited for professional video content such as training, explainer videos, and internal communications. It offers: A library of over 160 realistic avatar; support for more than 130 languages; custom avatars based on user-uploaded videos; integrated AI video editor for comprehensive video creation
BIGVU specializes in creating ultra-realistic talking avatars that can replicate your appearance and voice. It is particularly useful for: creating videos in over 29 languages; combining real and avatar videos with teleprompter support; adding branding elements and music to videos
D-ID provides advanced technology for creating interactive digital avatars from images and videos. Key features include: Real-time video creation with interactive agents; personalized adaptation and voice cloning; video translation and personalized video campaigns
Elai focuses on business presentations and training content, offering: over 80 avatars and 450 voice options; customizable presentation slides within videos; interactive in-video elements like quizzes