OpenAI launches data partnership for training AI models

New Delhi (Uttam Hindu News): OpenAI, run by Sam Altman, has launched a data partnership, where it will work with organizations to generate public and private datasets for training AI models. The company aims to collaborate with organizations to help AI models understand all subject matter, industries, cultures and languages, which requires the widest training dataset possible.

“The data partnership aims to help more organizations pioneer the future of AI and benefit from models that are more useful to them, including the content that interests them,” the company said in a statement.

ChatGPT developers said they are interested in large-scale datasets that reflect human societies and that are not already readily available online to the public.

“We can work with any method, including text, images, audio or video,” the company said. “We are specifically looking for data that expresses human intent in any language, subject, and format.”

OpenAI said it can work with data in almost any form and uses next-generation in-house AI technology to help people digitize and structure their data.

“For example, we have world-class optical character recognition (OCR) technology to digitize files like PDF, and automatic speech recognition (ASR) technology to transcribe spoken words,” the company said.

The company is looking for partners to help build an open-source dataset for training language models.

“This dataset will be public for anyone to use in AI model training,” OpenAI said. We’ll also explore using it to securely train additional open-source models. “We believe that open-source plays an important role in the ecosystem.”

“We are also building private datasets for training proprietary AI models, including our foundation models and fine-tuned and custom models,” it added.