Adobe has announced the release of two new versions of its text-to-image generative AI model, alongside new features for Firefly and updates to Photoshop and Illustrator.
The fourth-generation Firefly Image models follow the trend set by companies like OpenAI and Google with their chatbot releases, offering users choices between models that prioritize speed and efficiency or models that better respond to more complex requirements .
See also: Adobe patches 11 vulnerabilities in ColdFusion
According to Adobe, Firefly Image Model 4 is the fastest, most controllable, and most realistic Firefly model to date. It allows for the creation of images up to 2K resolution, offering more control over elements such as style, size, and shooting angle. The upgrades aim to improve image quality without sacrificing production speed.
For those who need even greater detail and realism, Adobe introduced Firefly Image Model 4 Ultra, capable of rendering complex scenes.
The new models are already available in the Firefly web app, along with other tools such as text-to-video and text-to-vector models, which are in public testing.
Additionally, the new Firefly Boards tool was introduced , a generative AI moodboarding app (similar to FigJam ), which was originally revealed as a Project Concept at the Adobe Max event in October. It is now in public beta. Adobe also announced that it is also preparing a Firefly mobile app for iOS and Android.
See also: Adobe releases Photoshop app for iPhone
Now, through the Firefly web app, users can also access AI models from other companies, such as OpenAI's GPT image model or Imagen 3 for image creation, as well as Veo 2 Google's for video. Support for other tools such as Luma, Pika, Runway, fal.ai and Ideogram.
Adobe clarifies that the third-party AI tools it provides through its platform are intended for experimental use and not for producing or publishing professional works. At the same time, it emphasizes that its own AI models are suitable for commercial use, which is not necessarily the case for the models from OpenAI, Google and Runway, as they are not trained exclusively on public or licensed content like Adobe's counterparts.
Additionally, as we mentioned earlier, Adobe has several new updates for its Creative Cloud. In Illustrator, generative shape fill and text to pattern tools are now widely available, having initially been introduced in a beta version. In Photoshop, features have been added that allow users to make intelligent color changes and automatically select details, such as hair, clothing or facial features.
Finally, Photoshop's Actions panel has been upgraded to offer editing suggestions tailored to the creator's personal style , and this feature is expected to form the basis for an AI creative assistant , which Adobe plans to integrate into its platform in the future.
See also: Adobe Firefly Bulk Create: Edit 10,000 images with one click!
Adobe ’s latest updates clearly show that the company is moving very aggressively in the AI space, trying to combine innovation with professional credibility . The fact that it emphasizes that its own models are “commercially safe” is strategically smart, especially in an era when the provenance and legitimacy of AI model training material is under scrutiny. However, making third-party tools available for “experimentation” is also a smart move, because it gives users freedom, but also keeps their distance when it comes to professional use – thus protecting both Adobe and the user. Overall, these developments show that Adobe wants to not just follow developments in AI, but to lead with reliable and professional tools.
Source: www.theverge.com
