Kneschke v. LAION (appeal)

AI training and copyright

Can AI be trained on protected works?

Closer to freely allowed

Appeal dismissed: dataset creation covered by the research exception

Holding

The court confirmed that LAION, as a non-commercial research association, could rely on the scientific text-and-data-mining exception, and held that an opt-out written only in natural language was not machine-readable and so ineffective for commercial mining. A further appeal to the Federal Court of Justice was allowed.

To opt out of AI training in Germany, use a machine-readable reservation, not just words in your terms.

Principles it bears on

• SupportsUsing protected works to train AI can be lawful under an exception or fair use.
• SupportsRights holders can reserve their works from AI training, but the opt-out must be machine-readable.

Sources

No official text linked yet: this summary relies on a secondary source. Verify against the original before citing.