Cornelia Funke kritisiert Künstliche Intelligenz: The Growing Battle Over “Intellectual and Creative Theft”
The tension between the world of high literature and the rapid expansion of generative artificial intelligence has reached a new flashpoint. During a recent public appearance, renowned author Cornelia Funke leveled a stinging critique against the current state of AI development, characterizing the unauthorized use of creative works to train large language models as “intellectual and creative theft.”
Funke, whose fantasy novels have reached millions of readers worldwide, is not alone in her apprehension. Her comments highlight a deepening rift between the tech industry’s drive for massive datasets and the creative class’s demand for agency, consent, and compensation. As Cornelia Funke kritisiert Künstliche Intelligenz, she is tapping into a global debate that is currently moving through courts and parliaments alike.
The remarks were made on Wednesday afternoon, April 15, during a session of the “Couchgespräche” hosted by Markus Loth at the St. Anna retirement home. The setting—a quiet, intimate dialogue—contrasted sharply with the weight of her message: that without clear regulatory guardrails, the very essence of human creativity is at risk of being exploited by machines.
The Core of the Conflict: Data Scraping vs. Creative Ownership
At the heart of Funke’s argument is the mechanism by which generative AI models function. To produce coherent prose, these systems undergo “training” on trillions of words scraped from the internet, including copyrighted books, essays, and articles. For many creators, this process feels less like technological progress and more like a massive, uncompensated extraction of human labor.
When authors speak of “theft,” they are referring to the fact that their life’s work is often ingested into training sets without their permission and without any royalty structure in place. This creates a paradox where the AI can mimic the style and nuance of a specific author, potentially competing with the very person it learned from.
This issue is currently being tested in multiple jurisdictions. High-profile legal battles involving major media organizations and tech giants are seeking to define whether “fair use” applies to the training of AI models. For authors like Funke, the lack of a “consent-first” model is a fundamental violation of intellectual property rights.
A Comparative Look: Human Authorship vs. Generative AI
To understand why this debate is so polarized, It’s helpful to look at the fundamental differences in how content is produced and how it is treated under current legal frameworks.
| Feature | Human Authorship | Generative AI |
|---|---|---|
| Source Material | Lived experience, emotion, and intentional study. | Statistical patterns derived from massive datasets. |
| Legal Protection | Protected under established copyright laws. | Legal status of AI-generated output is currently unsettled. |
| Economic Model | Direct compensation through sales, and royalties. | Data is often treated as a free public resource for training. |
| Intent | Narrative purpose and emotional connection. | Probabilistic prediction of the next token in a sequence. |
The Push for Regulation: The Role of the EU AI Act
Funke’s call for “clear rules” aligns with the legislative efforts currently underway in Europe. The EU AI Act represents one of the most significant attempts to date to regulate artificial intelligence by categorizing its uses by risk level and mandating transparency.
For the creative industries, the most critical components of such regulation are transparency requirements. Under proposed frameworks, AI developers may be required to disclose the copyrighted data used to train their models. This would provide authors with the necessary information to seek compensation or, at the very least, to opt out of future training cycles. However, the effectiveness of these “opt-out” mechanisms remains a subject of intense scrutiny among tech policy experts.
The struggle is not merely about money; it is about the integrity of the creative process. As generative tools become more integrated into writing workflows, the distinction between “human-made” and “machine-assisted” becomes increasingly blurred, complicating how we value and protect original thought.
Why the Human Element Still Matters
Despite the impressive capabilities of modern LLMs, the critique from authors like Funke suggests that there is a qualitative threshold that machines have yet to cross. The argument is that art is not just a sequence of likely words, but a reflection of human consciousness and cultural context—elements that cannot be “scraped” or “calculated.”
As the industry moves forward, the outcome of these debates will likely determine the economic viability of professional writing. If the “theft” Funke describes is allowed to continue unchecked, the incentive to produce original, high-quality literature may diminish, fundamentally altering the cultural landscape.
The next major checkpoint in this struggle will be the implementation phases of new digital copyright directives and the outcome of pending litigation in major tech-centric courts. These developments will decide whether the future of creativity is one of collaboration or one of extraction.
What do you think about the balance between AI innovation and copyright protection? Share your thoughts in the comments below and join the conversation.
Disclaimer: This article is for informational purposes only and does not constitute legal advice regarding intellectual property or copyright law.
