We’re introducing OpenAI Data Partnerships, where we’ll work along with organizations to supply private and non-private datasets for training AI models.
Modern AI technology learns skills and facets of our world — of individuals, our motivations, interactions, and the way in which we communicate — by making sense of the info on which it’s trained. To ultimately make AGI that’s protected and useful to all of humanity, we’d like AI models to deeply understand all issues, industries, cultures, and languages, which requires as broad a training dataset as possible.
Including your content could make AI models more helpful to you by increasing their understanding of your domain. We’re already working with many partners who’re desirous to represent data from their country or industry. For instance, we recently partnered with the Icelandic Government and Miðeind ehf to enhance GPT-4’s ability to talk Icelandic by integrating their curated datasets. We also partnered with non-profit organization Free Law Project, which goals to democratize access to legal understanding by including their large collection of legal documents in AI training. We all know there could also be many more who also wish to contribute to the longer term of AI research while discovering the potential of their unique data.
Data Partnerships are intended to enable more organizations to assist steer the longer term of AI and profit from models which might be more useful to them, by including content they care about.