Partager:

JAKARTA - The company behind ChatGPT, OpenAI introduced OpenAI Data Partnerships, a collaboration with organizations to generate public and personal datasets to train artificial intelligence (AI) models.

"To create a safe and beneficial AGI for all mankind, we want the AI model to understand in depth all subject matter, industry, culture and language, thus requiring the widest possible collection of training data," the company wrote in an announcement on its official website, quoted Friday, November 10.

The company claims to have made many partnerships, one of which is with the Icelandic Government to improve its GPT-4 ability in speaking Icelandic language by integrating their selected datasets.

Announcing OpenAI Data Partnerships — help steer the future of AI by collaborating on public and private datasets with us. https://t.co/4tbi5SZ6sS

"We are also partnering with the nonprofit Free Law Project, which aims to democratize access to legal understanding by including their many collections of legal documents in AI training," the company added.

This intended Data Partnership will allow the company's AI model to have broad information and insight, and allow more organizations to help direct AI's future and benefit from more useful models.

In this collaboration, OpenAI said it is interested in large-scale data collections that reflect human society and which are currently not easily accessible to the public, either in the form of text, images, audio, or videos.

How to partner with OpenAI

Currently OpenAI has two ways to partner, and may develop in the future:

Open Source Archives: That is, this dataset can be seen by anyone for use in AI model training.

Personal Data Collection: OpenAI also prepares a personal data collection to train its own AI model, including the basic model and model that has been adjusted and adjusted.

"If you have data you want to keep secret, but you want our AI model to have a better understanding of your domain, here's an optimal way to partner."

OpenAI also promises to treat your data with the sensitivity and access control level you want.


The English, Chinese, Japanese, Arabic, and French versions are automatically generated by the AI. So there may still be inaccuracies in translating, please always see Indonesian as our main language. (system supported by DigitalSiber.id)