If you're a creator with a stash of videos and photos, Adobe may be willing to cut you a nice little check. The software giant is going to use these assets and large data to train its artificial intelligence models. Adobe understands that in today's AI age, large data is like digital gold, and they are willing to pay handsomely for the raw materials they need to get the job done.
According to Bloomberg, Adobe is offering anywhere from 6 cents to 16 cents per photo, and an average of $2.62 per minute for videos. Depending on the contents of the video, the payment can go up to $7.25 per minute.
Adobe isn't just looking for any old footage. They're particularly interested in short clips that capture everyday human experiences and emotions - things like walking, laughing, crying, getting angry, and so on. The company says this type of content will be crucial as they work to develop a new text-to-video generator, which they plan to release later this year.
This move by Adobe is part of a broader push to expand their AI-powered tools beyond just image and illustration generation. By compensating creators for their work, they're hoping to build and own a strong dataset that will give their upcoming video tech a serious leg up on the competition. Although, there is no current text-to-video offering, available for public, Adobe is aiming to compete directly with models like Sora by Microsoft-backed OpenAI, that was announced a couple of months ago.
The use of copyrighted content to train large language models has been a controversial issue. Just last month, a group of authors filed a lawsuit against Nvidia, alleging the company used their copyrighted books without permission to train its AI platform. And in December 2023, The New York Times Company sued Microsoft and OpenAI, claiming they illegally used the newspaper's content to train their AI models.
This legal scrutiny around training data has clearly made some companies wary. In a recent interview, OpenAI's Chief Technology Officer Mira Murati admitted she wasn't even sure if the company's Sora model had been trained on user-generated videos from platforms like YouTube and Facebook.
Adobe seems to be taking a different approach with their AI efforts. By actively compensating creators for the videos and images they provide, the company is trying to build its AI capabilities in a more transparent way. Rather than risk legal headaches down the line, Adobe is upfront about how it plans to use the user submitted content. This proactive, creator-centric stance could give Adobe a real advantage. It's a smart move that could pay dividends later on.