Adobe’s VP of Generative AI, Alexandru Costin: 5 Key Questions Answered!

At Adobe MAX 2025, Adobe chose to develop and expand its in-house AI models, named Firefly. In its 24th edition, the company unveiled the fifth iteration of its image-specific model and also presented numerous partner AIs from companies like Google and OpenAI, thus creating a comprehensive catalog for its users. Alexandru Costin, Vice President of Generative AI at the company, explained this strategy in depth.

Adobe’s mission is to “become the company that powers creativity with AI”. How does generative AI concretely change the way creators interact with tools like Photoshop or Illustrator?

Our goal is to make AI both accessible and useful to everyone while preserving the creative integrity of our users. AI in tools like Photoshop or Illustrator doesn’t replace human creativity; it aids it. By integrating our Firefly model, we enable users to create images and videos more quickly and intuitively, giving them complete control over their creations. It’s not just about enhancing productivity; it’s about unlocking new creative possibilities.

Adobe has announced significant advancements with the Firefly models. What makes the Firefly Image 5 model stand out?

The Firefly Image 5 model features incredibly realistic high-resolution rendering capabilities, especially for generating human images, which is a significant challenge for AI. This model also includes powerful editing tools, such as photo restoration, allowing users to repair old or damaged photos. It’s a highly versatile model that opens up numerous possibilities for professional creators, whether they are photographers, illustrators, or designers.

Partnering with external companies is a core part of Adobe’s strategy. Why is it crucial for you to merge your models with those from companies like Google, OpenAI, or Topaz?

Incorporating third-party models allows our users to access the best tools available on the market. Each partner model has its unique features; for example, Google contributes its expertise in video, while OpenAI offers its linguistic capabilities. By merging these with our internal models, we enhance the user experience, thereby providing a wider variety of creative opportunities. This synergy keeps Adobe at the forefront of innovation and meets the diverse needs of our users.

How do you select the partner models that you integrate into the Firefly platform and your software, and what added value does this bring to your users?

We employ a rigorous scientific evaluation process to select models for integration. This process involves a team of researchers from prestigious universities who have developed scientific methods to assess the quality of models based on our clients’ needs.

Whenever a new partner submits a model, we evaluate it based on technical quality, professional utility, and its applicability to real creators. This ensures that we incorporate high-quality models that complement and enhance our offerings. For instance, with partners like Google, OpenAI, and Topaz, we bring unique capabilities, whether in image generation or video quality, while focusing on the creative experience of our users.

With generative AI continuously evolving, how do you see the future of human creativity in the AI era? Is there a risk of the dehumanization of creation?

AI is a tool, and like any tool, it doesn’t replace humans but extends their capabilities. The idea is to enable creators to focus on what they do best—their vision and artistic sensitivity—while leaving more technical or repetitive tasks to AI. We strongly believe that in the long run, AI will enhance human creativity rather than dehumanize it. This will allow more people to create, and those who are already creators to push the boundaries of their art. AI’s role is to amplify possibilities, always serving personal expression.