If you create content with artificial intelligence in Germany, a new law requires you to label it clearly — or risk fines. Here's what you need to know.
From August 2nd, if you've been using AI for your content creation in Germany, there's a new piece of legislation you need to know about.
It's called the KI-MIG (short for Künstliche-Intelligenz-Marktüberwachungsgesetz, or "Artificial Intelligence Market Surveillance Act") and it's Germany's enforcement of the EU's new AI Act. It outlines the rules that apply, as well as the institutions that will check whether you're complying and decide how much it'll cost you if you're not.
What do content creators need to know?
The part that matters most if you're a content creator: you need to clearly label AI-generated content as such. This applies to all professional, commercial or business activity.
So monetized influencers, digital marketers, publishers, agencies, and businesses must label all AI-generated content going forward. Non-monetised accounts, however, are exempt.
A few other caveats worth noting:
- AI-generated text does not need to be labelled if it has undergone human editing and review, and a person takes legal and editorial responsibility for the publication.
- Artistic and satirical content is also allowed some accommodations. But deepfakes require obvious disclosure.
- Explicit labelling is not required if content is "obviously" AI-generated or animated.
According to the new rules, posting AI content that is not labelled - which is currently widespread across the internet - can result in fines of up to €15 million, or three percent of global turnover for any business entities.
However, genuinely manipulative uses of AI carry fines up to €35 million, or seven percent of turnover.
Note that the fines above are the maximum penalties for prohibited AI practices according to the EU AI Act. Fines would be scaled down proportionally for small businesses and individuals.
Smaller slip-ups, like sloppy technical documentation for example, top out at €50,000.
READ ALSO: Explained - The EU's new rules on AI, deepfakes and chatbots
You might find it somewhat reassuring that this is not Germany "gold-plating" EU rules with extra bureaucracy. Berlin has explicitly avoided adding new obligations on top of what the AI Act already requires.
So, if you're already compliant with the EU rules, which officially entered into force on August 1st, 2024, you shouldn't have anything extra to do -- beyond ensuring that AI-generated content is labelled going forward.
If you do get caught out, you're not without recourse. You can challenge fines via Germany's standard administrative offence procedure through the courts.
If you're an employee who spots your employer misusing AI, you're now covered under Germany's whistle-blower protection law for reporting it.
Who's going to be checking up?
For content creators and most other people, it's the Bundesnetzagentur (BNetzA), Germany's default AI regulator that is tasked with making sure the updated rules are obeyed.
Other entities, such as state institutions, or transport and financial services companies answer to different entities — but if you're not sure whether you fall under them, the BNetzA central complaints office will point you in the right direction.
