OpenAI's head of Codex says AI still can't get creative design right: 'Give it up for the human brain'

AI is still learning.

The technology is revolutionizing all kinds of industries, but there are some things it still can’t quite get right, and, by most accounts, design is one of them. The gap is not just technical. It has to do with taste, and the judgment required to know when something actually works.

OpenAI’s head of Codex, Andrew Ambrosinoput it simply: Design is harder to measure than code.

“I think design’s a little bit harder to grade than software,” Ambrosino said on a recent episode of “Lenny’s Podcast.” “In that, creating a loop where you can train the model on what’s good design and what’s bad design is just a little bit more tedious and onerous than, you know, does the code compile?”

Anyone working in a creative field is likely worried about the impact AI will have on their careers. So far, it seems to serve as a useful tool that speeds up some processes, but it rarely produces a polished finished product on its own.

Dylan Field, the CEO of Figmaa digital design company that has taken on design behemoths like Adobe, has long argued that AI won’t replace human designers. AI models are trained on the “distribution of data” and typically create designs that people recognize as “average,” he said on a recent episode of “Hard Fork.”

The good news is that means creative jobs are safer for now. AI just means creatives have to learn new skills, like prompting or vibecoding.

As the Grammy-nominated musician Bas put it at a conference at Harvard in 2024: “At the end of the day, there’s going to be a certain level of human taste that’s still required to prompt the AI to create something that we’ll all enjoy,” he said.

“Let’s give it up for the human brain for now,” Ambrosino said.