AI’s Threat to Creative Culture: A Call for European Action
Generative AI is reshaping publishing, translation, journalism and other creative fields. The European Writers’ Council warns that the current approach favours model training over the people who produce culture. If policy does not change, European creativity risks being commodified, flattened and outsourced to opaque systems.
The Silent Plunder: How AI Exploits Creative Work
Many AI systems are trained on vast amounts of copyrighted material without consent, often under broad Text and Data Mining exceptions. Creators report lost income as platforms surface AI-written content that competes directly with human work. Cases include fraudulent AI books sold under authors’ names on major marketplaces and alarming reports of physical book destruction to feed scraping operations. Translators and niche-language authors face particular harm as models replicate their labour without remuneration, causing de-skilling and shrinking markets.
Beyond Algorithms: Protecting Human Creativity
The response must be practical and rights-based. The EWC calls for mandatory, human-readable labelling of all AI-generated content so audiences can tell machine outputs from human originals. Creators should be able to negotiate voluntary, individual licences for use of their works, and markets should account for the real economic impact of model training. An EU-wide, independent study is needed to quantify market harm, measure job displacement and map threats to linguistic and cultural diversity.
Shaping a Human-First AI Future
Europe’s strength is its cultural diversity and living creative workforce. Policy must prioritise originators: protect working conditions, guarantee social protections for creative labour and preserve the right to refuse AI replication of a work. Transparency, fair remuneration and solid evidence on harms are the pillars of a human-first approach. The choice is clear. Europe can protect its cultural legacy and set a global example by putting creators before models.
