From now on, developers in the EU will have to disclose the data sources used to train AI if copyright applies to that data.

When developers work on artificial intelligence, they need large amounts of data to train their AI models. To do this, they are increasingly using existing data that can be found online. However, much of this information is protected by copyright.

This happened with ChatGPT, for example, with visible consequences. For example, this chatbot can deliver text based on the work of a specific author. That is why the EU has now decided that in such cases developers must state exactly which data they used.

Revision AI Law

The decision came in the form of an addendum to the AI ​​Act. This forthcoming European directive is one of the first of its kind in which a government will arm this technology with the necessary legislation. The original text of the law was drafted more than two years ago, but a final text is still being worked on. This is a good indicator for this new decision.

However, an initial proposal to completely ban anything copyrighted for use in AI was rejected on the grounds of transparency. Fred Havemeyer, a software analyst at Australian group Macquarie, told Reuters that the European approach is tactful and puts the institute at the forefront of regulations surrounding AI technology.