How GPT Image 2.0 Redefines AI’s Role in Creative Work
Leonardo’s GPT Image 2.0 is charting a new course for AI in creative industries, according to Dwayne Koh, the company's Head of Creative. Unlike previous iterations of image generation models, GPT Image 2.0 isn’t just about technical precision—it’s stepping into the realm of creative partnership. From generating aesthetic ad campaigns to conceptualizing product ideas, this technology signals a major shift in how creatives approach their work.
Koh tested the model using straightforward prompts, avoiding the complex "prompt engineering" techniques often used to maximize AI outputs. The approach was deliberate: to see if GPT Image 2.0 could deliver professional-grade visuals with minimal input—exactly how marketers or designers would use it. The results were telling.
Creative Coherence Over Technical Perfection
Traditional benchmarks for AI image models often focus on technical details like rendering hands or crowded scenes accurately. But Koh argues these metrics miss the point. "The real test," he says, "is whether someone with no technical skills can input a simple idea and get a visually compelling result." With GPT Image 2.0, the generated outputs didn’t just meet technical standards—they displayed cultural awareness and creative coherence.
For example, when tasked with creating an ad for a Gen Z-targeted lip balm, the model added a "viral on TikTok" sticker—a feature not included in the prompt but one that resonated immediately as an effective marketing touch. This ability to go beyond mere image generation and suggest creative ideas is what sets GPT Image 2.0 apart.
Raising the Bar in Advertising and Beyond
Koh put GPT Image 2.0 to the test in advertising, a field where visual taste, audience understanding, and conceptual strength are critical. The model delivered results that rivaled outputs from professional campaigns, with thoughtful art direction and cohesive aesthetics. A fictional fashion brand ad, for instance, featured elements like smoke bombs and dynamic poses that matched the high-energy vibe of brands like adidas Originals or Vans.
However, its potential extends beyond advertising. Koh notes that GPT Image 2.0 can also be used as a product discovery tool. By generating a series of visuals for hypothetical brands or products, creatives can identify recurring patterns and reverse-engineer them into fully formed designs, identities, and market strategies. This "outputs-first" approach flips the traditional creative process on its head, allowing creators to test and refine ideas faster than ever.
Broader Implications for Creative Industries
The evolution of AI tools like GPT Image 2.0 shifts the focus from technical accuracy to "training taste." According to Koh, this means the models are increasingly capable of reflecting cultural context, suggesting trade-offs, and introducing new ideas—elements that have historically been the domain of human creatives.
These advancements could redefine workflows across industries. Instead of beginning with an idea and hoping it resonates, brands can start with AI-generated outputs, identify what works, and reverse-engineer concepts from there. This has significant implications not only for advertising but also for product development and branding.
What’s Next?
AI tools are no longer just accelerators of existing workflows; they’re becoming collaborators in the creative process. As models like GPT Image 2.0 continue to develop, they’ll likely push boundaries further, enabling marketers, designers, and product teams to experiment and innovate in ways previously unimaginable. The question isn’t whether AI can create—it’s how far it can go in shaping the future of creativity.