OpenAI Models Reportedly Trained on Diverse Web Data, Secures New Licensing Agreements

OpenAI's AI models have reportedly utilized data from 55 websites, encompassing businesses, non-profits, and government entities like the CDC and SEC. Concurrently, the company has finalized new content licensing deals, including one with News Corp., to broaden its training datasets. This development highlights ongoing efforts to expand AI model capabilities while raising questions about data sourcing and intellectual property.