Adobe launches its own AI image generator. The company announced a “family of creative generative AI models” called Adobe Firefly and releases the first two tools that take advantage of them. One of the tools works like DALL-E or Midjourney, allowing users to input a prompt and receive an image generated in response. Another generates stylized text, like WordArt, based on artificial intelligence.

At launch, Adobe is calling Firefly a beta version, and it will only be available through the website. But over time, Adobe plans to tightly integrate generative AI tools with Photoshop, Illustrator, and Premiere.

Adobe Inc. Chief Product Officer Scott Belsky told Bloomberg that the company has never used artist content to train generative AI.

Earlier, independent software developers Krita Foundation discovered a clause in Adobe’s terms of service that refers to the use of AI for content analysis. Some artists have interpreted this as the company’s use of images to train generators like the DALL-E 2 and Midjourney.

Belsky called the criticism “a wake-up call” and said the company uses AI to analyze products and improve their features.

“We are rolling out a new evolution of this policy that is more specific,” Belsky said.

He added that if the company ever decides to develop generative AI, Adobe will announce it publicly.

One way Adobe hopes to stop AI from using other people’s artwork is to offer artists a way to block AI from learning on their artwork. The company is working on a Do Not Train system that would allow artists to embed a request in an image’s metadata that could stop learning systems from viewing it.