Adobe has pledged not to use customer data to train its AI models, a stance aimed at reassuring creative professionals who rely on its tools. The commitment comes as Adobe integrates generative AI features into products like Photoshop, Illustrator, and Premiere Pro under its Firefly brand, which includes models trained only on content Adobe has legal rights to use.
Adobe’s promise contrasts with other AI companies that use publicly scraped data, leading some creators to worry their intellectual property may lose value. Adobe chief strategy officer Scott Belsky told Axios, “You have to be explicit not just about what you are going to do, but what you aren’t going to do,” highlighting the importance of transparency in data use policies.
Adobe recently codified this commitment in response to concerns over changes to its terms of service, which raised questions about potential shifts in data practices. To further support transparency, Adobe allows creators to digitally sign their work and specify if they want it used in AI training.
However, Adobe does use content from its Adobe Stock contributors to train Firefly and provides a “contributor bonus” for such usage. While this move aligns with Adobe’s legal and transparent approach, the company’s adoption of generative AI remains controversial among artists who view the technology as a potential threat to their careers.
