Sora Access Halted After Artist Protests
OpenAI has temporarily suspended access to its text-to-video tool, Sora, following protests from artists. The backlash centers on concerns over copyright, intellectual property, and the ethical use of generative video technology. The incident highlights growing tension between creative acceleration and the need for responsible AI implementation, particularly regarding attribution and transparency for source training data.
- A group of early testers and creative partners, under the name 'Sora PR Puppets', intentionally leaked a version of Sora on the platform Hugging Face as a form of protest. This leaked version allowed users to generate 10-second, 1080p videos from text prompts. - The protesting artists claim that hundreds of creatives provided unpaid labor for OpenAI, a company valued at $150 billion, by conducting bug testing, offering feedback, and doing experimental work. OpenAI stated that participation in the testing and feedback program was voluntary. - A central complaint from the artists is the restrictive nature of the early access program, which required all generated content to be approved by OpenAI before being shared. This led the artists to believe the program was more focused on public relations and advertising than on genuine creative exploration and critique. - There is significant controversy surrounding the data used to train Sora. OpenAI's CTO, Mira Murati, has been unable to confirm whether the model was trained on videos from platforms like YouTube, Facebook, or Instagram, stating only that it used "publicly available and licensed data". - YouTube's CEO, Neal Mohan, has stated that if Sora used YouTube content for training, it would be a "clear violation" of the platform's terms of service, which prohibit unauthorized scraping or downloading of content. - The protest is not against AI-generated art itself, but rather against OpenAI's specific business practices. The artists' open letter even recommended open-source alternatives for AI video generation. - This incident is part of a broader legal and ethical debate surrounding AI and copyright. OpenAI is already facing multiple lawsuits for alleged improper use of text from books and news articles to train its other AI models. - The Sora controversy has also raised concerns about the generation of deepfakes and the unauthorized use of likenesses, with instances of AI-generated videos featuring deceased celebrities like Robin Williams causing distress to their families.
