Decisions · Court · Germany

GEMA v. OpenAI

Landgericht München I (Regional Court) · 2025-11-11 · First instance · Ongoing

Later historyUnder appeal.

LG München I, 42 O 14139/24

Liability for AI outputs

Is an AI provider liable when outputs copy protected works?

Provider liable

Memorising song lyrics in the model infringed copyright

Why hereCoded at the pole: its core holding or recommendation answers the question outright.

The model as a copy

Is a trained model a copy of its training works?

Closer to a copy of its training works

Lyrics memorised in the model are reproduced in it

Why hereCloser to 'a copy of its training works': the court found the memorised lyrics reproduced in the model itself, not only in outputs, and read fixation broadly (a probabilistic, lossy representation from which simple prompts recover the text is enough). The finding covers the memorised works, not every training work.

Holding

The court found that song lyrics memorised in OpenAI's models and reproduced in outputs infringed the rights GEMA represents; the text-and-data-mining exception did not cover this. The judgment is under appeal.

In Germany, a model that memorises and reproduces protected works can infringe, even if training itself might be allowed.

Principles it bears on

SupportsAn AI system that memorises or reproduces protected works in its outputs can infringe.

Sources

Primary source ↗Summary ↗

This record

Data as of 2026-10-21.

Drafted with AI assistance and checked against the sources linked above; not yet reviewed by a lawyer. Methodology

Changes

2026-10-20Also coded on the new question on the model as a copy.