Adobe is paying $3 a minute for AI-training video of people touching things
Adobe is pushing its way into the growing business of generative AI video, joining OpenAI’s Sora and Google’s Imagen 2.
The new tools will roll out this year, according to Adobe.
In contrast to its web-scraping rivals, Bloomberg reported that Adobe is paying videographers up to $120 for stock footage used to train the model.
High-priority subjects include: footage of people showing emotions, clips of people touching things, and anatomy shots of eyes, hands, and feet.
AI companies are growing increasingly wary of copyright lawsuits, as giants like YouTube threaten possible litigation if AI is trained on their videos. Plus: AI is learning so fast that the data used to train it could be completely tapped by 2026.
High-priority subjects include: footage of people showing emotions, clips of people touching things, and anatomy shots of eyes, hands, and feet.
AI companies are growing increasingly wary of copyright lawsuits, as giants like YouTube threaten possible litigation if AI is trained on their videos. Plus: AI is learning so fast that the data used to train it could be completely tapped by 2026.
