EU's New AI Labeling Rules Kick In: What Changes for Businesses on August 2
Published on 08/04/2026 at 05:23 | Redaktion boerse-global.de
The European Union's artificial intelligence rulebook has moved from paper to practice. As of August 2, organizations across the bloc face binding transparency obligations when deploying AI systems that interact with people or generate content. The requirements, rooted in the EU's AI Act, touch everything from customer-service chatbots to synthetic media.
Machines Must Now Identify Themselves
Article 50 of the regulation forms the backbone of the new duties. Any AI system engaging directly with humans—think recruitment tools or virtual assistants in support centers—must clearly disclose its machine nature. Users should never be left guessing whether they're talking to a person or an algorithm.
The labeling mandate extends to deepfakes as well. Realistic audio, imagery, or video depicting individuals or events now requires visible marking. AI-written text touching on matters of public interest also carries disclosure duties, unless a human editor has exercised editorial oversight. Purely private use, along with artistic, satirical, or fictional output, escapes the requirement.
Fines That Bite
Non-compliance carries a heavy price tag. Companies face penalties up to €15 million or 3% of global annual turnover for transparency violations. The ceiling climbs to €35 million or 7% of worldwide revenue for prohibited AI practices.
Existing systems get breathing room until December 2 to retrofit compliance measures—machine-readable watermarks or technical metadata, for instance. High-risk AI applications enjoy extended timelines, with some deadlines stretching into December 2027 or August 2028.
New regulatory duties can feel overwhelming, but staying compliant doesn't have to be. Just as the AI Act demands transparency, UK employers face their own legal obligations under the Health & Safety at Work Act 1974—and many unknowingly fall short. A free toolkit with 9 ready-to-use tools, including risk assessments and checklists, helps you identify gaps and protect your workforce. Download the free Health & Safety at Work Act 1974 Toolkit
Industry Voices Frustration Over Missing Guidance
The effective date has arrived, but practical clarity hasn't. The eco industry association warns that harmonized technical standards and European Commission guidelines failed to materialize in time, leaving businesses in what it calls a "regulatory blind flight." Bitkom echoes the concern, pointing to widespread uncertainty even though one in three German companies already uses AI.
Germany has designated the Federal Network Agency (Bundesnetzagentur) as its supervisory authority. Austria, meanwhile, has yet to name a national watchdog. Some organizations aren't waiting for further direction—the Austrian broadcaster ORF has already pledged to timestamp AI usage in its productions.
Training Workers Becomes a Compliance Pillar
Beyond labeling, the AI Act pushes companies to build internal competence. Article 4 obliges employers to foster employee understanding of AI systems—how they function and where risks lie. The goal: a workforce that can critically assess the technology rather than passively rely on it.
Building workforce competence is a theme that extends beyond AI—it's central to workplace safety too. Over 37,000 UK businesses already use a free Health & Safety Toolkit with practical templates, checklists, and toolbox talks to train staff and meet their legal duties. Whether it's fire safety, PPE, or first aid, everything you need is ready for immediate use. Download the free Health & Safety Toolkit
Adoption rates vary across the region. Around 30% of Austrian companies use AI, compared with roughly 20% across the EU. More than 80% of AI-adopting firms already run training programs. For management, the takeaway is straightforward: staff education isn't just good practice anymore—it's a regulatory requirement that also supports operational efficiency.
