Adobe's approach, while potentially costly, could enable the sourcing of higher quality footage and allow for customization of the video generation tool. The company is cautious about copyright issues, a problem that has embroiled OpenAI in numerous lawsuits. When Adobe ran a similar program for its Firefly AI image generator, it did not accept photos of brands, public figures, or specific words, and it is expected to apply similar rules to its video generator program.
Key takeaways:
- Adobe is developing an AI model that can convert text into video and plans to pay creators for the material used to train the AI.
- Unlike OpenAI, which has been accused of using copyrighted works for training, Adobe plans to purchase footage from filmmakers for training its AI model.
- The company is reportedly buying 100 different types of footage, with compensation for submitted footage ranging from $2.62 to $7.25 per minute, with each video being worth $120.
- Adobe's approach, while potentially more expensive, could allow the company to source higher quality footage and customize the upcoming video generation tool to its liking.