You’ve probably heard of the EU’s sweeping copyright reforms—but did you know one article could determine whether your content is fair game for AI training? Article 4 of the DSM Directive governs text and data mining (TDM), and unless you explicitly opt out, your work could be used without your knowledge or consent. This guide explains what Article 4 means, how it affects you as a writer or publisher, and exactly how to protect your rights.
“Opt-out is your right. Silence is interpreted as consent.”
Use both legal language and technical methods to defend your work from unauthorized exploitation.
What is Article 4?
Article 4 of the EU Directive on Copyright in the Digital Single Market (Directive (EU) 2019/790) governs Text and Data Mining (TDM) by organizations across the EU.
TDM is the automated process of analyzing digital content to extract data patterns—a technique used heavily by AI companies to train language models.
What Does It Say?
-
Research exception: Article 4 permits TDM for commercial or non-commercial purposes unless the rights holder has expressly reserved their rights.
-
In other words: AI developers can mine your public content by default unless you actively opt out.
How to Opt Out
To reserve your rights under Article 4, you must clearly express this via:
-
Machine-readable means (e.g., your
robots.txtfile) -
Human-readable form (e.g., your website's Terms of Use or Legal Disclaimer)
Once declared, companies that continue scraping your content may be in breach of the Directive and subject to legal enforcement in EU member states.
Example Robots.txt Snippet
This snippet blocks general crawlers and AI bots from accessing and indexing your site.
User-agent: *
Disallow: /
User-agent: GPTBot
Disallow: /
User-agent: Google-Extended
Disallow: /
Where It Applies
-
Applies across all EU Member States
-
Enforceable through national copyright laws
-
Includes protections for both text and data sets (e.g., metadata, structured listings)
Why It Matters
If you don't opt out, your articles, images, recipes, guides, or metadata can be legally used by companies building AI models—without notifying you, crediting you, or compensating you.
By invoking your rights under Article 4, you:
-
Retain control over how your content is used
-
Join a growing collective of rights holders pushing back
-
Help shape responsible digital standards in Europe
Need Help?
Visit www.eatw.org or contact us at