OpenAI has launched Sora, a new AI video generator, allowing premium ChatGPT users to create videos from text prompts. However, the company has restricted the ability to generate human likenesses to a select group of invited testers to mitigate concerns over misuse, such as deepfakes and misappropriation of likeness. The tool has produced creative examples like videos of sumo-wrestling bears and a cat drinking coffee. OpenAI emphasizes its commitment to blocking inappropriate content, including nudity and harmful material. The initial response to Sora was so overwhelming that OpenAI has temporarily halted new account creations. The company aims to engage with artists and policymakers before wider public release. Additionally, OpenAI has not disclosed the specific sources used to train Sora, which raises questions about copyright practices. The tool’s launch follows previous scrutiny over OpenAI’s use of copyrighted materials in its training datasets. Overall, Sora represents a significant advancement in AI-generated media while navigating ethical considerations.
