The Question Europe's Art World Is Actually Asking About AI

A panel discussion scheduled for September 9 at Berlin Art Week 2026 carries a title that deserves unpacking: "AI-Based Art in Europe, Regulate, Support, Enable?" The framing suggests three distinct policy options. But are they actually in tension? And if so, where exactly does the disagreement lie?

The event, hosted by ZKM Karlsruhe at the Baden-Württemberg State Representation in Berlin, brings together artists Tristan Schulze and Mario Klingemann (known professionally as Quasimondo), alongside Sven Meyer from Staatsschauspiel Stuttgart and Tina Lorenz, who leads artistic research at ZKM. The panel composition itself signals something worth noting: this is not a debate between technologists and humanists, or between regulators and creators. Everyone on stage works at the intersection of art and technology. The disagreement, if there is one, must be more granular.

Three Words, Four Different Conversations

The title's three verbs deserve disaggregation. When someone says "regulate AI art," they might mean: (a) copyright rules for training data, (b) disclosure requirements for AI-generated content, (c) restrictions on deepfakes and synthetic media, or (d) classification standards for what counts as "AI art" in funding applications. These are four different regulatory conversations with different stakeholders, different trade-offs, and different urgency levels.

Similarly, "support" could mean: direct grants to artists working with AI, institutional investment in AI research labs like ZKM's Hertzlab, tax incentives for galleries showing AI work, or educational programs training the next generation of media artists. "Enable" might refer to: access to compute resources, legal clarity that removes uncertainty, or simply cultural legitimacy that makes AI art fundable and exhibitable.

Until these terms get disaggregated, the conversation risks becoming a performance of positions rather than an actual negotiation of trade-offs. The question worth asking: which of these regulatory, support, or enabling mechanisms is most contested, and why?

The Institutional Landscape

ZKM Karlsruhe occupies a particular position in this debate. As the panel description notes, the institution approaches AI "both as a creative tool and as an object of critical reflection." This dual stance is not neutral. It represents a specific bet: that artistic practice and critical inquiry can coexist within the same institutional frame, and that this coexistence produces better outcomes than separating "makers" from "critics."

This model has precedents. Art Laboratory Berlin's recent workshop with artist Helena Nikonole demonstrated what critical artistic practice looks like in action. Nikonole's methodology involves "intentionally pushing AI systems beyond their intended applications to reveal their ideological underpinnings and structural limitations." This is neither pure celebration nor pure critique. It is a form of inquiry that uses artistic practice as a research method.

The question for policymakers: does this hybrid model scale? ZKM has decades of institutional history and stable funding. Can the same approach work for independent artists, smaller institutions, or emerging practitioners who lack access to comparable resources?

What the Artists Actually Need

The panel includes two artists whose work represents different approaches to AI. Tristan Schulze has participated in discussions at institutions including Luxembourg Art Week and ZKM, engaging with questions about AI's impact on creative practice. Mario Klingemann, working under the name Quasimondo, has been a prominent figure in generative art for years, predating the current wave of large language models and diffusion systems.

Their presence raises a question that policy discussions often elide: what do artists working with AI actually need from institutions and regulators? The answer is unlikely to be uniform. An artist using AI as a tool for generating visual elements has different needs than one using AI as a subject of critical inquiry. An artist training custom models has different compute requirements than one using commercial APIs. An artist selling work through galleries faces different copyright questions than one creating public installations.

The strongest version of the "regulate" argument would acknowledge that some regulation protects artists. Clear copyright rules, for instance, could benefit artists whose work is used to train models without compensation. The strongest version of the "enable" argument would acknowledge that excessive regulation creates barriers to entry, particularly for artists without institutional backing or legal resources.

The Institutional Perspective

Sven Meyer's presence on the panel, representing Staatsschauspiel Stuttgart, signals that the conversation extends beyond visual arts. Theater institutions face their own questions about AI: synthetic voices, AI-assisted scriptwriting, digital scenography, and the automation of production processes. The trade-offs differ from those in visual arts, but the underlying tension remains similar.

Cultural institutions across Europe are navigating a period of uncertainty. The EU AI Act creates new compliance requirements, but its application to artistic practice remains unclear in many cases. Funding bodies are developing criteria for AI-related projects, but these criteria vary across jurisdictions and programs. The S+T+ARTS program and similar EU initiatives have created frameworks for art-science-technology collaboration, but access to these programs requires institutional capacity that many smaller organizations lack.

The Question Behind the Question

The panel's framing assumes that "regulate," "support," and "enable" are the relevant policy categories. But this framing may itself be contested. An alternative framing might ask: who benefits from the current uncertainty, and who is harmed by it?

Large technology companies benefit from unclear copyright rules that allow training on vast datasets without compensation. Established institutions benefit from funding structures that favor organizations with track records in AI art. Artists with technical skills benefit from a landscape where AI literacy creates competitive advantage.

Conversely, artists whose work is used for training without consent are harmed by regulatory ambiguity. Emerging practitioners are harmed by funding structures that favor established players. Audiences are potentially harmed by a lack of transparency about what is AI-generated and what is not.

This reframing does not resolve the debate, but it clarifies what is at stake. The question is not simply whether to regulate, support, or enable, but whose interests each policy choice serves.

What Would Have to Be True

For the "regulate first" position to be correct, it would have to be true that the harms from unregulated AI art outweigh the costs of regulatory compliance, and that regulation can be designed in ways that do not disproportionately burden smaller actors.

For the "support first" position to be correct, it would have to be true that the primary barrier to AI art is resources rather than legal uncertainty, and that public investment can be allocated in ways that do not simply reinforce existing institutional hierarchies.

For the "enable first" position to be correct, it would have to be true that the primary barrier is cultural or legal legitimacy rather than resources, and that removing barriers will benefit a broad range of practitioners rather than primarily those already positioned to take advantage.

The Berlin Art Week panel may not resolve these questions. But if it succeeds in disaggregating them, in moving from a debate about labels to a conversation about specific mechanisms and their trade-offs, it will have accomplished something valuable.

The conversation continues across European institutions grappling with similar questions. For those tracking how these debates evolve, the Human × AI Content Hub offers ongoing coverage where cultural policy meets the broader European AI landscape.