18 September 2026
On 2 August 2026, the European Union began actively enforcing key parts of its AI Act. If your business licenses AI tools, uses generative AI in your marketing or product development, or is watching how AI regulation is likely to evolve here in Australia, this is worth understanding — even though it's EU law.
Let's break it down.
What Actually Started on 2 August
Two things came into force. First, transparency obligations under Article 50 now apply to any AI system that talks to a person or produces synthetic content — chatbots must disclose that users are dealing with AI, and AI-generated content (including deepfakes) must carry machine-readable marks identifying it as artificially generated. Existing systems have until 2 December 2026 to comply with the marking requirement.
Second, and arguably more significant for the AI industry, the European Commission's AI Office now has active enforcement powers over providers of general-purpose AI models — the large models that power tools many Australian businesses already use. Providers must document their models, publish summaries of the material used for training, and maintain a policy for complying with EU copyright law. Penalties for non-compliance reach €15 million or 3% of global annual turnover, whichever is higher.
Why This Reaches Beyond Europe
The AI Act applies to any provider whose AI systems are placed on the EU market, or whose outputs are used within the EU — which in practice means most major AI model providers, regardless of where they're headquartered, now have binding obligations to disclose what their models were trained on and how copyright compliance is handled.
For Australian businesses, this has a few practical implications:
- You may start seeing more disclosure from your AI vendors about training data sources and copyright policies, simply because providers now need this documentation to operate in the EU — and it's often easier for a vendor to apply one global standard than maintain different versions for different markets.
- If your business licenses content, brand assets, or creative work to AI providers, understanding how those training-data disclosure obligations work gives you a clearer picture of what's actually happening with your material once it's used to train a model.
- This is a preview of where regulation is heading. Similar transparency and disclosure expectations are increasingly part of the global conversation around AI and copyright, including here in Australia — so understanding the EU framework now gives you a head start on compliance expectations that may eventually apply closer to home.
What to Watch
The high-risk AI obligations under the EU AI Act — the rules governing AI used in sensitive areas like employment, credit, and law enforcement — have actually been delayed to December 2027 and August 2028. So this isn't the full regime landing all at once. It's a narrower, more immediate set of transparency and disclosure rules, with the bigger obligations still to come.
How IP Solved Can Help
Whether you're licensing AI tools, considering how your own content and brand assets interact with AI training, or just trying to understand what these global shifts mean for your business, IP Solved can help you make sense of the IP angle.
Get in touch with IP Solved today to discuss how AI regulation developments — in Europe and closer to home — might affect your business.
This article provides general information only and is not legal advice. Specific advice should be obtained for your business and target markets.
