Responses to the EU's AI Strategy for the Cultural and Creative Sectors consultation

In July 2026, the European Commission opened a call for evidence on an AI Strategy for the Cultural and Creative Sectors – one of the flagship actions in its Culture Compass for Europe. The call sought views on how artificial intelligence is affecting creation, preservation and access to culture across the EU, and on how a future strategy should protect cultural and linguistic diversity and creators' rights while supporting innovation. The call was open to anyone: individual citizens, businesses, professional associations and public bodies. It closed on 11 September 2026, and its findings will feed into a strategy due in the first half of 2027.

ITI responded to the call because translation has a longer history of engagement with AI than most of the sectors this strategy will cover. Machine translation and, more recently, generative AI have been present in the language services sector for years, which means that translators can offer the Commission evidence about the impact of sustained technological disruption, rather than speculation.

ITI's submission argued that AI should complement rather than replace human judgement, and that a strategy focused on increasing the volume of machine-generated content will not necessarily achieve the hoped for cultural and linguistic diversity.

Who responded

The Commission received 5,375 submissions to its call for evidence. Of these, 126 mentioned translation and were submitted by individual translators, national and European professional associations, businesses and public bodies across Europe and beyond.

Most submissions came from individual citizens (81 of the 126, including 21 who identified themselves as working translators). The remainder came from organisations, including several sector bodies: CEATL (European Council of Literary Translators' Associations), FIT Europe, ATRAE (the Spanish Association for Audiovisual Translators), AVTE (AudioVisual Translators Europe), the Austrian Forum for Literary Translation, and ITI.

What they said

Unsustainable working conditions

Pay and working conditions came up repeatedly, particularly in submissions from sector bodies. ATRAE wrote: "Professionals are struggling, rates have plummeted, conditions have worsened and many find the current situation unsustainable." It said machine-translation post-editing (MTPE) had been used for years to make work "more precarious". De Creatieve Coalitie, which represents cultural and creative workers in the Netherlands, cited its own survey of 713 workers published in December 2025. It reported that approximately one in three translators had experienced fewer assignments or lower income attributable to generative AI. Individual translators described the same pattern.

"Finally, the implementation of AI along with post-edition further widen [sic] precarious conditions in an already mistreated profession." Anonymous EU citizen

Data used for training LLMs without consent

Individuals and organisations alike raised the concern that AI systems were trained on translators' and authors' work without permission or payment. ATRAE called for AI companies to be transparent about the datasets used to train their models and to work on an opt-in consent basis.

"[W]e know for a fact that the same AI that replaced us and that now translates everything was precisely trained using our previous (human) translations. Which, of course, they didn't pay us for. They stole our work, our expertise and our creativity, and used it to create a tool made exactly to replace us and kick us out of the market."
Cristina Domínguez, translator

Loss of skilled work, not just jobs

Several submissions described post-editing as a change in the nature of the work as well as in its pay. FIT Europe argued that translation should be treated as an early-warning case for other cultural and creative professions. CEATL said that using AI to pre-translate a text "cuts human translators out of the first draft, which is essential to the success of literary translation". AVÜ said fees must rise "in order to allow for the acquisition and maintenance of human expertise in the language professions".

"At the same time, translators are increasingly asked not to translate texts themselves but to post-edit or repair machine-generated output. Many find this work less professionally satisfying and less creatively fulfilling, while it can also be associated with lower remuneration and greater pressure to work quickly." FIT Europe

More output is not the same as more diversity

Several submissions argued that a larger volume of multilingual content does not amount to greater linguistic diversity. CEATL said the promise that AI can open up minoritised languages through dominant "bridge" languages "contravenes authors' rights and leads to literary, linguistic and cultural impoverishment". AVÜ warned that widespread use of AI risks homogenisation that harms Europe's linguistic and cultural diversity.

"A Europe in which enormous quantities of inexpensive machine-generated multilingual content circulate, while fewer professional translators can earn a sustainable living, is not necessarily a more linguistically diverse Europe." FIT Europe

A call for firmer rules

The submissions did not all ask for the same thing. ATRAE called for a ban on generative AI in creative and cultural spaces. CEATL asked for full transparency for all models placed on the EU market. It also asked for public funding and awards to be made conditional on not using AI technologies that "exploit, pollute, and endanger creative labour and the environment". AVÜ proposed a specific rule on pay: "Post-editing of AI-generated language versions must not be remunerated less than translations produced by language experts themselves."

Where ITI's submission fits

ITI's submission made four main points:

• AI should complement rather than replace human creativity.
• Translation, with a longer history of engagement with AI than most cultural sectors, offers evidence of the effects of sustained technological change.
• More machine-generated multilingual content will not necessarily deliver greater linguistic diversity, which depends on the continued presence of skilled human professionals.
• The strategy should make the maintenance of skills and entry routes an explicit objective, rather than assuming that "upskilling" will cover it.

ITI also noted that as post-editing replaces translation from scratch, translators risk losing broader competence while facing downward pressure on rates.

Several of these points were echoed elsewhere. FIT Europe argued that linguistic diversity cannot be measured by the quantity of multilingual content. It warned that machine output may weaken "the human infrastructure on which meaningful linguistic diversity depends". CEATL and AVÜ also linked AI-generated translation to a loss of diversity. FIT Europe and AVÜ raised concerns about the replacement of skilled work and the maintenance of professional expertise. ATRAE, FIT Europe, CEATL and AVÜ all described falling rates and deteriorating conditions. ITI's position is consistent with the views put forward by these other translators' bodies.

A note on sources

Quotations are taken from the Commission's "Have your say" portal as published, including original spelling and punctuation. The sector bodies' supporting statements were not reviewed, only the text shown on the portal. Some submissions were written in other languages and appear on the portal as automatic English translations. The AVÜ text is one of these.