Skip to content
Skjld Labs

AI labelling becomes law in the EU: what Norwegian businesses should do now

From 2 August 2026, AI-generated content must be clearly labelled in the EU: deepfakes, AI-written text on matters of public interest, and chatbots that must disclose they are machines. Norway has no AI law yet, but the rules reach Norwegian businesses all the same. We walk through what the labelling duty actually requires, where Norway stands, and why building transparency into your systems now beats retrofitting it later.

3 min read Laget med AI-hjelp
Documents, images and videos flowing past the EU stars, marked with small symbols

On 2 August 2026, the transparency obligations of the EU's AI Act took effect: AI-generated content must be clearly labelled. The duty, recently covered by kode24, targets three categories: deepfakes, AI-generated or AI-manipulated text on matters of public interest, and interactive AI systems such as chatbots. Norway, meanwhile, still has no AI law of its own, and incorporation into the EEA Agreement remains unresolved. That combination creates a vacuum Norwegian businesses should navigate deliberately.

What the labelling duty actually requires

The regulation was adopted in 2024 and phases in until August 2028. The transparency requirements now in force boil down to three things. Synthetic content, whether images, video, audio or text, must carry machine-readable marking, for instance watermarks or other markers. Deepfakes must be visibly labelled for humans, not just machines. And systems that interact directly with people must make clear that the user is talking to a machine. A chatbot can no longer pass itself off as human.

The duty sits with whoever provides and deploys the systems, not the model vendor alone. A business publishing AI-generated content or operating a customer chatbot owns the responsibility itself. The EU has published a Code of Practice to make compliance concrete, and violations can draw substantial fines.

Norway lags behind, but the rules still bite

Norway has no AI law yet. The government plans a public consultation in autumn 2026 and a legislative proposal in spring 2027, with Nkom designated as the national coordinating AI supervisor. But two facts mean Norwegian businesses cannot sit back in the meantime.

First, the EU rules already apply to Norwegian systems used in the EU market. If your chatbot has users in Sweden or Denmark, or you publish AI-generated content reaching an EU audience, you are in practice covered, regardless of the pace of Norwegian legislation. Second, the delay is no free pass at home: privacy law, copyright law and the penal code apply to AI-generated content today, as the minister for digitalisation has also stressed.

What this means in practice

For businesses building or deploying AI systems, the requirements come down to concrete design choices. Chatbots and agents should identify themselves as AI from the first message. It is a small UI decision with major legal weight. Content production with AI in the value chain needs traceability: the system should know, and be able to document, which parts of an end product were generated, so labelling can happen automatically where the rules require it. And where images, audio or video are generated, machine-readable marking in the form of metadata and watermarking should be part of the production pipeline, not a manual afterthought.

It is a familiar lesson from privacy work: transparency built into the architecture costs little, while transparency retrofitted onto production systems costs a lot. Businesses that already log what is AI-generated and label it consistently avoid both the scramble when the Norwegian law lands and the risk in the interim.

Transparency as a competitive edge

It is tempting to read the labelling duty as yet another compliance burden. We think that reading is too narrow. Customers, employees and partners place more trust in businesses that are upfront about their use of AI, and systems designed for traceability are consistently easier to debug, evaluate and improve. Labelling then becomes a by-product of good system design rather than a chore.

For Norwegian businesses the advice is simple: don't wait for the 2027 proposal. Map where AI already produces content or meets customers today, make sure your AI systems can account for what they have generated, and let labelling follow from the architecture. Those who build this in now will meet the EU requirements, the coming Norwegian law and their own customers' expectations with calm.

From article to practice

Want to put this to work?

Local and Norwegian AI. By default, the solutions we build run on our own servers in Sandefjord. If your data is sensitive, we set them up on site at your premises.

Read about Local and Norwegian AI Book a free call
Safe and local AI

Read on.

Newsletter

AI in practice, one email a month.

What we've tested, one tip you can use, and what's coming next.

You'll get the Skjld Labs newsletter once a month, and can unsubscribe at any time. Privacy

First step

Thirty minutes. Free.

A no-obligation conversation about what you are facing. We will say so if AI is not the right answer. We are happy to meet you at Langes gate 8 in Sandefjord, or on video.