Since 2 August 2026, new transparency obligations under the EU Artificial Intelligence Act have applied to certain uses of AI-generated content. What does this mean for your company’s multilingual communication?

What came first, the chicken or the egg? Technology develops at an exponential rate, while regulation traditionally takes longer to design and implement. This creates the challenge of regulating AI uses that are already emerging while allowing innovation to continue. With the establishment of the EU AI Act, Europe has introduced one of the world’s first comprehensive legal frameworks for artificial intelligence. The Act takes a risk-based approach, setting different requirements for different types of systems and prohibiting certain practices. But what does this mean for companies using AI to produce multilingual content?

Since the arrival of AI tools, new use cases have emerged rapidly, along with new risks. The EU AI Act was adopted in 2024, and has been implemented progressively. A major milestone was reached on 2 August 2026, when the transparency obligations came into effect. As a language services provider of the translation and multimedia space and integrating AI, automation and human expertise into our localisation workflows, we want to explain what AI use in multilingual content workflows means from a legal, governance and quality perspective.

AI has changed how multilingual content is managed. Is AI Act compliance now changing the rules?

AI has made it possible to create multilingual workflows and automations that speed up website and e-commerce updates and enable faster information sharing. But as AI becomes more widely used, organisations also need to consider how these workflows fit within the EU AI Act and its risk-based approach.

Under Regulation (EU) 2024/1689, AI systems are classified according to the level of risk they pose, with different requirements applying to each category. Unacceptable-risk practices are prohibited, while high-risk systems are subject to stricter requirements because of the potential significant risks they pose to health, safety and fundamental rights. Other AI applications fall into lower-risk categories and generally face fewer demands. For multilingual workflows, transparency is particularly relevant. Article 50 introduces specific transparency obligations for certain AI systems and AI-generated content. These obligations came into effect on 2 August 2026. They include informing people when they are interacting directly with AI and ensuring that certain AI-generated or manipulated content can be identified.

What does the EU Artificial Intelligence Act say about AI-generated text?

The focus is therefore moving from “Can AI generate or translate this content?” to “What happens before we publish it?”. Under Article 50(4), organisations deploying AI systems that generate or manipulate text to inform on matters of public interest must disclose that the text has been artificially generated or manipulated. However, an exception applies where the content has undergone human review or editorial control and a natural or legal person holds editorial responsibility for the publication.

The crucial distinction: AI-generated content vs. human-reviewed content

This makes the meaning of “human review” particularly important. It is not simply a matter of clicking “approve”, running a spell-check or correcting grammar.

The European Commission clarifies that human review involves deliberate examination of the substance of the content by people with relevant knowledge and professional judgement. Editorial control also means having the authority to approve, alter or reject content on substantive grounds, including fact-checking and assessing the trustworthiness of sources. Superficial or purely procedural checks, such as spell-checking or grammatical correction, do not qualify as human review or editorial control. In other words, meaningful human judgement needs to be built into the workflow.

<h2>Human review: ensuring content reliability, accountability and AI Act compliance

Now that AI can accelerate content production, the key question is where meaningful human judgement enters the workflow. Beyond linguistic quality, human review plays a crucial role in verifying the reliability of AI-generated content, identifying misleading or inaccurate information, assessing the trustworthiness of sources and ensuring that the final message is appropriate for its intended audience.

For multilingual content, this means combining linguistic expertise with factual verification, subject-matter knowledge and editorial judgement. This is where AI governance becomes particularly relevant: organisations need to understand not only what AI produces, but also who verifies its output, who approves its publication and who ultimately takes responsibility for the information communicated across languages and markets.

AI disclosure: a risk or an opportunity for brands?

Does telling audiences that AI was used undermine trust or build it? Transparency can help build trust when audiences understand where AI enters the workflow and who reviews, amends, rejects or approves the final content. But there is more to consider.

AI watermarking is another development to watch. In August 2026, Anthropic announced that future Claude models would use text watermarking to help determine whether Claude was involved in generating text, including translations. The watermark is designed to be invisible to readers and does not change the content itself.

For businesses, this raises an important question: as AI-generated content becomes easier to identify and disclosure becomes mandatory, are transparency and human accountability becoming integral parts of content strategy?

Confused? Rely on the CPSL Hybrid Approach for AI compliance. Our experience with automation, machine translation post-editing (MTPE) and AI workflows, combined with human expertise and standards such as ISO 18587, helps organisations balance efficiency with quality, thorough oversight and accountability. As AI regulation continues to evolve, the right approach is to bring together innovation, appropriate governance and human expertise. Talk to us.