Adobe is reportedly buying video content (for around $3 to $7 per minute) to train Sora-like AI generator

[Roboticker]

 

Adobe, the maker of Photoshop and Premiere Pro, is reportedly acquiring videos to train its new AI tool, a text-to-video generator. The software giant’s venture into the AI-generated video content field is seen as a response to the growing demand for such technology. 

 

Designed to generate videos based on textual input, Adobe’s ambitious project aims to compete with OpenAI’s Sora. The company is leveraging its network of artists and photographers to source assets for AI training. However, the rates offered have been criticized as being relatively low, considering the potential value of the AI models.

 

The company’s approach involves offering artists around $3 (according to Bloomberg)—and up to $7.25—per minute of video to be used as training data. This strategy, however, has raised eyebrows as the artists are likely not going to be credited for their work nor receive subsequent royalties. The question of ethics in AI training data acquisition is thus brought to the forefront.

 

Adobe’s foray into AI-generated video content, while innovative, raises important questions about the value of creative work in the age of AI. How do you think artists should be compensated for their contributions to AI training? Share your thoughts below.

 

 

Read more:

• https://www.bloomberg.com/news/articles/2024-04-10/adobe-is-buying-video-clips-for-3-per-minute-to-build-ai-model
• https://www.pcmag.com/news/adobe-buying-videos-train-sora-competitor
• https://www.zdnet.com/article/adobe-is-buying-videos-to-train-its-new-ai-generator-but-are-they-paying-enough/

 

Image: J P | Dreamstime.com