OpenAI employees discussed book costs for early ChatGPT training: documents
by@WSJ
Summary
Recently unsealed documents in a copyright case reveal that OpenAI employees deliberated on the costs associated with purchasing books to train early ChatGPT models. This scrutiny into OpenAI's data sourcing practices underscores the company's evaluation of various options for acquiring large text datasets during the initial development of ChatGPT.
Analysis
OpenAI: OpenAI develops and deploys advanced artificial intelligence models and systems for language processing and other applications. The company created the ChatGPT series of conversational AI tools. It is currently facing a copyright lawsuit in which internal employee discussions about sourcing books for training early versions of these models have surfaced through unsealed court documents. Data Sourcing: Early development of ChatGPT involved evaluating options for obtaining large text datasets such as books. Legal Scrutiny: Unsealed documents in an ongoing copyright case have highlighted OpenAI's internal considerations on acquiring data for model training.
Categories
aimachine_learningtech