A deeper look into the EU text and data mining exceptions: harmonisation, data ownership, and the future of technology

AI training and copyright

Can AI be trained on protected works?

Freely allowed

Extracting information from works shouldn't need an exception at all

What it says

The authors find the EU exceptions too narrow in key ways and argue that extracting informational value from protected works should not require a text and data mining exception in the first place. They warn the current rules may push EU developers toward older, biased data or models trained abroad.

Two of Europe's best-known copyright scholars on why the EU approach may backfire.

Sources