News on Medial

Here's Proof You Can Train an AI Model Without Slurping Copyrighted Content

WiredWired · 8m
Here's Proof You Can Train an AI Model Without Slurping Copyrighted Content

French researchers have released a large AI training dataset composed entirely of text in the public domain, challenging the belief that copyrighted materials are necessary for training AI models. Meanwhile, non-profit organization Fairly Trained has awarded its first certification for a large language model built without copyright infringement. Developed by 273 Ventures, the KL3M model was trained on a curated dataset of legal, financial, and regulatory documents. Fairly Trained certifies companies that train their AI models using data they own, have licensed, or is in the public domain. The availability of infringement-free datasets like these could revolutionize the AI industry's reliance on copyrighted materials.

Comments

Download the medial app to read full posts, comements and news.