From PetaPixel:

Generative AI companies like OpenAI are skirting copyright rules by using unclear training methods for their models. OpenAI’s Sora, a text-to-video AI, generates lifelike videos. During an interview, tough questions were asked about Sora’s training, revealing issues with using licensed data or data from platforms like Shutterstock without clear consent. The debate continues on AI violating artists’ copyrights, with companies reluctant to discuss the issue openly.



Read more at PetaPixel: OpenAI’s CTO Won’t Discuss Training Data for AI Video Generator Sora