AI Training & Data Mining Policy ~ TorrentFreak

TorrentFreak welcomes readers, researchers, and journalists who wish to reference and build on our work. The text of our articles is published by TF Publishing under a Creative Commons Attribution-NonCommercial (CC BY-NC 3.0) license. That license permits reuse, with attribution, for non-commercial purposes.

The use of our content to train, fine-tune, or ground artificial intelligence systems is a separate matter, and this page sets out our position on it.

Training Requires a License

Using TorrentFreak content to develop commercial AI products is a commercial use. It falls outside the non-commercial scope of our Creative Commons license.

This includes, without limitation:

– Training or fine-tuning large language models or other machine-learning systems on our articles, archives, datasets, or metadata.
– Retrieval-augmented generation (RAG), semantic indexing, or any system that ingests our content to generate or ground responses in a commercial product.
– Text and data mining (TDM) of our website or feeds for any of the above.

Any commercial AI developer or platform wishing to use TorrentFreak content for these purposes must request permission and enter into a separate licensing agreement with us. The CC BY-NC license does not grant this right, and no such use is authorized without an executed license.

Reservation of Rights

TorrentFreak expressly reserves all rights in its content against text and data mining and AI training, including the machine-readable reservation of rights permitted under Article 4 of the EU Copyright Directive (and reflected in the EU AI Act’s transparency framework). Our presence in any public dataset, crawl, or index does not constitute consent.

The absence of a technical block does not imply permission. Equally, where we deploy technical measures, robots directives, or other signals to opt out of automated collection, those signals must be respected.

What We Offer

We are not opposed to AI. We are opposed to uncompensated, unattributed use of work that takes time, expertise, and original reporting to produce.

For AI developers who wish to license our archive and live feed, we are open to discussing the option to provide structured, machine-readable access on commercial terms.