Image Generation: Artificial intelligence, creativity and design - edited by Hamza Shaikh

See a preview of the book on Amazon. In a nutshell, I really liked this book, especially the final chapter, because, like Hamza, I'm an optimistic dreamer at heart.

There's something unsettling about watching an entire industry transform in real time.

I've been following AI image generation since Midjourney first appeared, and I'll admit—I was skeptical. Another tool promising to "revolutionize" design. But Hamza Shaikh's "Image Generation: Artificial intelligence, creativity and design" cuts through the hype and delivers something far more valuable: a practical guide grounded in actual use cases from practitioners who've been in the trenches.

What makes this book stand out is its hybrid approach. Shaikh doesn't position AI as either savior or threat. Instead, he argues that "AI is not a replacement for human creativity but an extension of it." That perspective runs throughout the book, and it's refreshing. No grandiose promises. No doomsday scenarios. Just honest exploration of what these tools can actually do.

The timing here matters. Much of the book focuses on developments from the past three years—the period when AI image generation moved from research labs into actual practice. Shaikh traces the arc from Midjourney's community-driven approach through Stable Diffusion's open-source release to ControlNet's introduction of image translation algorithms. These weren't just technical milestones. They fundamentally changed how designers work.

The book features contributions from many practitioners, including some I've connected with on LinkedIn who are actively working at this intersection:

• Carlos Banon
• Tim Fu
• Will Garner
• Omer Nuray
• Hassan Ragab
• Kier Regan-Alexander
• Ismael Seleit

Their chapters add real-world context to Shaikh's frameworks. You're not just reading theory—you're seeing how these tools function in actual design and construction workflows.

One section that particularly resonated explores how "over 80% of human perception is shaped by sight, making it our primary channel for absorbing information." In an age where attention is currency and doomscrolling is standard behavior, Shaikh argues that AI image generation isn't just faster rendering—it's a fundamental shift in how we iterate on ideas. The ability to generate "countless variations of a concept in moments, breaking down the barriers of time and manual effort" changes the design process itself.

The practical tutorials deserve mention. Real-Time Façade Iterations using ComfyUI and Grasshopper. Generative 2D to 3D workflows with TripoSR and Blender. AI Form-Morphing with Stable Diffusion. These aren't superficial overviews—they're detailed enough to actually implement. Shaikh's tip to "start with clear design intent" and never "expect AI to generate ideas from scratch" reflects hard-won experience, not marketing copy.

There's an inevitable gap here, and Shaikh would likely acknowledge it. The field moves at exponential speed. Developments from the past three months—tools like Kontext, Nano-Banana, and Seed Dream that specifically aid image generation for architects—couldn't make it into this edition. That's not a criticism. It's reality. Any book about AI is partially outdated the moment it's published.

But here's what doesn't become outdated: understanding how these systems actually work. Shaikh emphasizes that "learning how AI processes information and responds to different inputs is far more valuable than memorizing keywords." That insight will remain relevant regardless of which specific tools dominate next year.

The book also doesn't shy away from harder questions. The chapter on data center energy use and AI's climate impact. The honest assessment that "programs to generate entire detailed and custom floorplans, 3D-detailed façades and resolved structure solutions have not kept pace" with image generation capabilities. The acknowledgment that architects are often "burdened with debt from their education, engaging in internships or working overtime and underpaid."

If you want to understand where AI image generation currently stands—not where the hype says it's going, but where practitioners are actually using it today—this book delivers.

The YouTube interview with Shaikh is essential viewing for anyone serious about this space.

The real value here isn't predicting the future. It's giving you frameworks to adapt as that future unfolds. Because as Shaikh notes: "change has always been the only constant."