Every six months or so, a new AI tool drops and the internet loses its collective mind. "Graphic design is dead!" they cry, sharing screenshots of logos and posters that, frankly, look like they were designed by someone who once walked past a design studio. This week it's ChatGPT Images 2.0 prompting the latest round of panic. And once again, the panic is aimed at entirely the wrong thing.
What happened
OpenAI shipped a significant update to its image generation capabilities. ChatGPT Images 2.0 produces noticeably more polished visual output than its predecessor — cleaner layouts, better typography handling, more convincing compositions. Social media did what social media does: declared graphic design officially over.
Meanwhile, designer Jonny Burch published a refreshingly honest counterpoint. Despite using AI tools daily across his entire workflow, he's blunt: AI still can't design well without heavy human guidance. Left to its own devices, it produces what he calls "slop" — generic, tasteless output that wouldn't survive a first crit.
And then there's the quieter but arguably more important story. UX designers are increasingly being expected to deliver production-ready code and assets — not just wireframes and prototypes. The line between designer and developer is dissolving, and AI is the solvent.
Three stories. One week. And they're all pointing at the same thing.
The bigger picture
Here's what I think most people are getting wrong about this moment. The question isn't "will AI replace designers?" — it won't, not any time soon. The output still lacks taste, context, and strategic thinking. Anyone who's actually tried to ship AI-generated design work into a real project knows this.
But that's not the threat.
The real shift is subtler and already underway: AI is redefining what stakeholders, clients, and product teams expect designers to deliver. And that's a much harder problem to navigate than an existential one.
Think about it. When a product manager can generate a "good enough" mockup in ChatGPT in thirty seconds, the value of a static visual comp drops to near zero. When AI coding tools can scaffold a front-end from a design file, suddenly "why can't you just give us production-ready code?" becomes a reasonable question in a sprint planning meeting.
The bar for visual output is being commoditised. That doesn't mean design is dead — it means the definition of design work is mutating beneath our feet.
I've seen this play out in teams I've worked with. Designers who used to spend two days on high-fidelity mockups are now expected to spend that time on systems thinking, interaction logic, and user research synthesis instead. The craft hasn't gone away. It's just moved.
And honestly? That's not all bad. The parts of design that AI handles passably — generating layout variations, colour exploration, asset production — were never the interesting bits anyway. The interesting bits are the decisions. Why this layout and not that one. What happens when the user is angry, or confused, or in a rush. How this screen fits into a journey that started three touchpoints ago.