1. Adobe is developing an AI model for video generation but hasn’t provided a launch date or many details.
2. The model will be integrated into Premiere Pro, offering features like object addition, object removal, and generative extend.
3. Adobe is exploring collaborations with third-party vendors like OpenAI to enhance its video generation capabilities but these integrations are still in the early stages.
Adobe is developing an AI model to generate video that will be integrated into Premiere Pro later this year. This model is part of the Firefly family of generative AI products and will offer features like object addition, object removal, and generative extend. The model creates footage from prompts or reference images and aims to help users enhance their video editing tasks.
To address concerns about deepfakes, Adobe will introduce Content Credentials to indicate AI-generated media and the specific model used to create it. The company is also reportedly compensating photographers and artists on Adobe Stock for submitting short video clips to train its video generation model. This approach contrasts with rivals who are facing lawsuits for using copyrighted content without proper compensation or credit.
Customers will need generative credits to use the upcoming video generation features in Adobe Premiere, with allotments ranging from 25 to 1,000 per month depending on the subscription plan. Adobe is also exploring partnerships with third-party vendors, including OpenAI, to integrate their video generation models into Premiere. These collaborations are still in the early stages and not yet available to customers.
While Adobe’s efforts to enter the generative video space show promise, the lack of specific details and the underwhelming reputation of its image generation models raise questions about the future success of its video generation features. The company’s collaboration with OpenAI and other vendors indicates a commitment to innovation in generative AI, but the actual impact remains to be seen.