The company is reportedly paying artists on its stock media platform, Adobe Stock, up to $120 for short video clips to train its video generation model. This approach contrasts with rivals like OpenAI, which allegedly uses publicly available web data. Adobe is also considering integrating third-party video generation models into Premiere, including OpenAI's Sora. However, the company has not revealed the cost for customers to use these upcoming video generation features.
Key takeaways:
- Adobe is developing a generative AI model for video creation, which will be integrated into its Premiere Pro video editing suite. The model will power three new features: object addition, object removal, and generative extend.
- The company is also introducing Content Credentials to identify AI-generated media and the model used to generate it, in an effort to address concerns about deepfakes.
- Adobe is reportedly paying contributors on its stock media platform for short video clips to train its video generation model, a move that contrasts with rivals like OpenAI who have been accused of using publicly available web data without permission.
- Adobe is also in talks with third-party vendors, including OpenAI, Pika, and Runway, about integrating their video generation models into Premiere Pro.