Adobe sets tone for fair AI development by paying for AI training materials

Adobe, the software giant renowned for its creative tools, is embarking on a groundbreaking venture into the world of AI. The company has announced that it is building its own AI model capable of transforming text into video that will be seen as a direct competitor to the recently announced Sora AI model from OpenAI.

However, unlike its counterpart, Adobe is taking a different route by directly compensating creators for the material used to train its AI model. The move is a great example of fair and ethical AI development where content creators are acknowledged as the driving force for these models.

OpenAI has been in the spotlight countless times, first with ChatGPT and now with Sora, for sourcing its AI training materials from public places without acknowledging the original creators. The company is currently battling multiple lawsuits over copyright infringement.

By engaging with filmmakers to procure footage for AI training purposes, Adobe aims to ensure transparency, fairness, and quality in its AI-driven video generation process.

The footage sought by Adobe spans a diverse range of everyday scenarios, from mundane activities like using a phone to human emotions and interactions.

This comprehensive approach aims to capture the richness and complexity of human experiences, enabling the AI model to produce more nuanced and lifelike video content.

Adobe is purchasing 100 different types of footage. The report calculates that compensation for submitted footage ranges from $2.62 to $7.25 per minute, with each video being worth $120.

In line with its commitment to ethical standards, Adobe has established guidelines to govern the sourcing of training material. While the requirements have not been officially laid out they will expectedly be similar with a similar program Adobe ran to train its Firefly image generator.

Contributors were not allowed to send photos of brands, public figures, and even specific words.